Orthogonal for Mutual Conductance

Hello, and welcome to the macro fab engineering podcast. We're your hosts, Parker, Dolman and

Steven Craig. And this is episode 107.

Well, pos there, Steven. Yeah. Brewing.

Yeah, had a little bit of a mishap

actually was just looking at like, what this isn't like, oh, yeah, I know what these are.

So So we've actually covered the genesis of my electric brewing here on the, on the podcast. So I've gone from their drum solo. So I've gone from a USPS cardboard box, and I've upgraded to a to a legit junction box for for all of my electronics. But about two weeks ago, I actually melted my SSR in my box.

So I have to ask, because SSRS are like, their size of like a match book. About to match book stacks. Yeah, so the pretty big. How much smoke did that let out? I

caught it before. So here's the thing. It was functioning. I caught it before anything bad happened. I turned on the whole system. And it was heating up some water and everything was running fine. And I turned to my wife and I was like, Do you smell something and I looked over and there was just little wealth of smoke coming out the side of the bar. So I kicked it off. It was still running, it was still doing everything. Ah,

so seating up. The thing is, I

don't even know if I actually destroyed the SSR. It's just the plastic casing around it started to deform and the jacket of the wire that was connecting to it that melted. Oh, it got really hot. But it didn't actually stop. You still have it? No, actually throw it away. I got it, guys. I've already replaced it. Yeah, I

think we should should next time, because it won't be next time.

I hope not. I've mitigated the problem.

Oh yeah, you put a fan on it right

yet? Well, I bought I bought a little 12 volt blower actually, because it has a fan. A fan is a little annoying because it kind of has like a dead space right in the center cog. So I got a blower instead. And I got a blower that has the right dimensional cage fan blower right. Yeah, it has the right mouth dimensions that it was basically the same size of the heatsink. And I basically stuck that on the back and I ended up since that was 12 volts. And I only have 220 volts. In my case, I ended up buying a little switch mode, DIN rail guy and slepping that in there, and I fixed the whole thing for like 20 bucks, total, and, you know, an afternoon worth of drilling through the case. So the thing is like that SSR that I had. It's been running for I don't know what a year and a half, maybe even a little bit longer. Yeah. And it's it's done a lot of brew days. And the thing is, I always had it passively cooled. I had an on a heat sink with compound heating compound, but it was in a box that was generally sealed. It had a couple holes in it. I never had any forced air cooling in it. So I'm not really that surprised that it finally gave up the smoke.

Yeah. And when you said because I I'm using the same SSRS for mine. Yeah. And so my solution was I just ordered extras, because I'm designing my enclosure to be as waterproof as possible. So I don't want to have like a fan hole in the side. Gotcha. So I might have to paralyze them or yeah, we'll see what happens.

Well, I had their heatsink. This SSR is rated for 40 amps at 220. And I never pushed it past 25 amps. Yeah, so you know, derailed with heat, heat and all that stuff. But now that I have that fan blowing on it or the blower, continuously, the air if you feel it coming off of it doesn't even it's not even warm. It's still cold. Yeah. And so so it works out really well. And I gave it a test the other day I put it I'd had a 10 hour brew day and probably five of those 10 hours it was Ronnie constant. Yeah, constant in some form. Maybe full power, half power, whatever. And no problems whatsoever. So yeah, that was uh, that was kind of fun.

So next time that happens, either one of mine or yours, whatever. Yeah, yeah, we pull it out, put it aside, you know, put a new one in finished brew day. And then we take that module and then get it the go and just see how much smoke that thing let's just let it choose. Just yeah, just go nuts. Yeah, see how much it will actually like, you know?

Well, from what I've researched online, those style of SSRS like the brick wall. Yeah. They're they're unfortunate. Failure Mode is to fail, closed, fail and then dump as much as they possibly can. The good thing is like, on the back end of my I'm not using it to limit power if it goes 100% I really don't care. Because I have a 5500 element wad element on the backend, which it's normal for that to go full

power and that element is in five to eight gallons of water. Yeah, so as long as you don't like decide to like, go across usin and then come back. Oh, I never leave the pot. Yeah, never leave it so like because only the worst thing could happen would be that pot evaporated six seven gallons of water in your apartment. Oh yeah. Imagine how swampy you would be in your apartment if that happened

actually, so no, it's funny. I gal I boiled off one gallon of water actually measured it one gallon of water in one hour on my last brew day. Because I had I had it on a rip roaring Yeah,

your humidity in your apartment and from 10% to 80 Oh,

no, no, no, it was 100 looked in the kitchen and like you could barely see the ceiling because it was just like a fog of all

tropical forests and

with that I'd open a window and like suck it all out with a fan. So I need to find a better way I need like an exhaust hood. Yeah, eventually where this brew thing will land which is likely a garage one day, I will put a proper because because actually the ceiling. You can see like water spots on the ceiling. Yeah, it started not. I mean, it's it's ceiling material, I guess what? Gypsum or whatever. And so like, it's fine, but it was like you can see little damp spots on this. Another landlord's gonna come in and be like,

What the hell just gets in trouble tropical plants and vines. Yeah. Monkeys. Leopards. Tropical kitchen. Yeah, kitchen be like Rainforest Cafe.

Random monkeys.

animatronics that are really bad. Yeah.

Yeah, that's a really really inefficient way of making humidity in a place just boil the snot out of water. And because of this, so because of this SSR melting, I'm trying to take safety a little bit more seriously, just a fraction. I mean,

didn't you get a more proper electrical box to uh,

I just made mine a little more proper. Okay, I went out and started looking for more proper ones. And then I found some on Amazon. I just I'm not ready to plop down that money. Yeah, especially because it works right now.

Cuz you got a like in wall it's not a breaker box. But it's like a junction box. Yeah,

it will it it started its life as a as a breaker box. Okay, I got it all the breaker like terminals and stuff. And then just made all my own cracks. The one I've got is like a It's one forum machine uses an external use. Yeah, yeah. Yeah. Yeah. And so it's got a gasket and everything. I want to get that one. That one looks really nice. It's just, I, I don't my blue rig doesn't warrant the size of that case yet. Yep. And I want to plan out like all the electronics in there. And so it was just, I figured, instead of spending like 300 bucks and getting all the extra electronics, I'll spend 20 bucks and fix this thing. Yep.

So what are you gonna do to fix it?

Well, first of all, I mean, I've ran all the wiring. Yeah. to the to the larger gauge. That's a SSR Yeah, larger gauge and I kind of double stranded just more juge. Yep. But I haven't bought this yet. It's gonna be a better one to buy a GFCI for

it. interrupt your quick church. Yeah. So if people that watch AV E. It's like something's working. I guess that's the best way to explain it. Yeah, there's gonna be a that's a Canadian term, though. There's gonna be a Texan term for that

fix into fixing. Yeah, fixing is a total tax. Yeah, yeah. That's not the same as church though. It's right. I'm just saying like, that's like, yeah, fix into does not mean that you are repairing or fixing that's true. It means you're about to do something. Yes. Or you are planning on doing something. Or you even have a thought to do something like it means Yeah, it's goofy like that. So okay, well, for one thing, if something fix it, you could actually say, Yeah, that would work. That's valid English here in Texas. Yeah. Perfect. So yeah, no, I want to add a GFCI. breaker to this box because I am dealing with liquids. Yep. And a whole lot of power. And

so I want to know how you're implementing that because I started implementing GF. Ci to my setup to Okay, so yeah, there's been You haven't even seen what I'm trying to do.

Well, okay. So just for those who don't know, a GFCI breaker is not a regular breaker, correct. A regular breaker will just once you trip or go beyond what the maximum current that breaker set for it pops and opens up. Yep. A GFCI monitors how much current flows out of one leg and monitors how much current flows in the other leg and they must match within a small portion or a small amount. Yes. So typically that trip point is like five milliamps. And so what would basically is there to guarantee is that like kickoff current law, or Yeah,

it doesn't decide to jump into you.

Or if it does decide to jump into you, it will limit it to five milliamps. Yeah, I'm sure five milliamps at 220 volts hurts like hell yeah,

but that's not I don't think that's enough to kill you.

It is not. Yeah. No, it's not. So the thing is right now, I don't have I have a standard breaker on my system. But you know, if for some reason, everything were to go catastrophically wrong. They would happily dump 50 amps through me through you in into the floor. And yeah, right. Right. Right. But I mean, then again, so many other appliances in your house. Yes. So, you know, usually GFCI is are only in your bathroom, and sometimes your kitchen. You know, like, you got a hairdryer. And you drop it in the sink kind of thing that's usually on a GFCI. So you don't have a lot of hair dryers have a built into their plugs to write that little bit. A little plug that has the button that says test on it. That's a GFCI. Yeah. And so because I'm working with water, if something were to happen, I know like, what if I spilled it on the box? And then it was all over me and things like, so I want to get GFCIs but they're unbelievably expensive. Yeah, tally up the entire cost of the box with all the electronics in it. That's cheaper than just one. GFCI. Yep. And the GFCI because and my box, I want it to be able to handle 50 amps. No, yeah, a 50. Amp 220 volt GFCI. It's like 300 bucks. Yep. They are not cheap. That's what I found, too. Yeah. So one solution that you can do is put the GFCIs downstream and put them so that adjust on your element. Yeah. So that way you walk every five. Yeah, yeah, you only have to do 25 amps. You could have an individual GFCI for each element. And then your cost goes down to like 80 bucks per Mm hmm. So that's a solution. But then all the rest of the electronics are not GFCI controlled. The thing is like, you have to balance your load. So if you have if you're running like a 120 volt thing on with other 220. circuitry, that doesn't work with a GFCI, because you're pulling more from one leg than the other. Yep. So you kind of have to have a GFCI for each individual thing. And

so, this so my setup is all to 20. Yes, I don't have any 120 a month. So all the load should be balanced. Yeah. And so I went out and bought a 50 amp 240 volt spa GF GFCI. Oh, nice. It's a little box. It's designed to go right on your spa. Okay. And I basically, it the 80 bucks, nice. So it has the metal box and everything. And I basically opened it up and just gutted it and got I took the breaker out and the mount for it. And that's going in my box. And I'm just throwing out the little tiny metal box that came with.

You know, I was thinking about doing something similar to that having the GFCI for the whole system be a separate box that's on the outside. Yeah, such power goes into that box, then power leaves that box and goes into like, command center box. Yes.

I'm just putting that thing in my command center box. Gotcha. I can just mount it. Yeah. So that's what I was gonna try to do. That's

a good idea. Yeah. I know that spas a very similar kind of Yeah,

environment, except that you're not cooking yourself.

Yeah. What is the heating element? In a spa? Is it is it directly connected to the water?

Wouldn't surprise me. So is or is there a heat exchanger?

I don't know. A heat exchanger. It probably is a little much. Yeah, I'd

probably really expand. It's probably just I bet you it's a water tank heater. Yeah. Just like It's like no, it's probably like a rim setup where it's like it's in a tube. And you just pump water into the tube.

Maybe but you can you can heat up. You can have it just going without a pump going. Right. You know, eating without a pump now. Okay. Okay, so it probably is. Yeah, it probably just flows water over a heating element. Yeah, I bet you it's a like a $12 heating element from Home Depot.

And so in some spas are like gas powered. So. Okay. Yeah. Just depends. So yeah. So how are you going to the GFCI?

I'm not sure on that.

I'll send you that link. Because that way you can just run it on your 20s Well, that's it depends because you were running for wire. 220 That's right, cuz I'm only running three wire, which is why I have to go everything to 20.

Right. Well, because you have two hearts in a neutral, right? You

don't have a ground. Now I have a ground. Two hots and a ground. Yeah. Okay. Yeah. The 220 legs that come into your house? Yeah. Okay, I got basically I'm using my welder plug as my brooder plug, and that's only a three prong so you got you got to 20 You know, one phase ground and and to 20 Inverse phase is what you have. I'm sort

of doing the same thing as you because I don't even really use my neutral.

They don't. Well, you have your pumps are 120

Oh, yeah, but right now I'm not my pumps. I'm doing them to the wall.

Oh, then yeah, just use this thing because only 80 bucks you can run everything on your setup. As long as it works. I don't know if it works yet. It should.

I'll have to check that. Yeah. Cool. All right. So Parker. Yeah. Talk to me. What's up?

Zack board? We're talking about a couple weeks ago. The digital analog converter for the Raspberry Pi. Mm hmm. Got it gotten built? Tested. Oh, very cool. Yeah. So first, I've I've seen it. Yeah. So this that one that you have there? What's it say? It says that's Rev. One. See? That's the cheap one. Oh, yeah. Yep. Yeah, this is the expensive one. So you can see how its bypass caps are ginormous. And fill me look. Yes, point one microfarad film caps. And then I've got the B rev, which has the cheap bypass caps, and then the fancy resistors. And capacitors.

So all this, all this capacitor jazz up at the top? Is that power supply? Yep. Filtering. Okay. Yeah. That's a pretty decent amount of filtering on that.

Yeah, I did the typical what I usually normally do for LBOs, which is like one mic 100. No, I use 1010 microfarad. Point one microfarad. And then point one mic, Fred. 10, microfarad. And then, depending on the application, I think these are like I think those are 10, microfarad electrolytics. To always put unless you're lytic. Usually, they're two, because they just a big bulk cap. Yeah. And they have different frequency responses. Yeah. So that's the that's the trick when you pro tip is when you're looking at your filtering on power supplies, fish like this. And because you always want to put a couple capacitors, usually. And you want to make sure that their frequency responses are different. And you don't want them because they'll have a notch somewhere in their frequency response. And you don't want the notches lined up. Because if you have the, you know, if the notches are in the same spot, then they don't filter your your filtering doesn't filter that spot.

Because it does but just not as good as it could. Yeah, yeah, yeah,

I think the rule of thumb is divide by 10. Right. So whatever, you have a large capacitance, like if you have a 10 mic, divide by 10, you can put a one mic down, and then divide by 10. You keep if you want to have as well, the thing evens curves that all line up. The trick with that is you have to make sure that the same package and the same dielectric, that's the only reason. That's the only way that rule of thumb lines up, right? Because if you change the package, or you change the dielectric that shifts that whole setup,

right, because the reason why the curve even curves upwards, is self inductance. That happens later off, which depends on you know, geometry and a handful of other things. Yeah, package. But what you got going on here, a lot of times these electrolytic caps have a high ESR they were high, comparatively ESR ceramics, yeah, you can put your bolt capacitance in the aluminum electrolytic and then put a ceramic next to it and the ceramic will take, you know, the real fast stuff. Yep. Whereas the electrolytic takes the bulk of the just like stream of power.

Yeah. So cool. It's you gotta think of as the ceramics are more of a high pass style. filtering for your power supply and electrolytics are more of a low passer. Kind of,

I guess. Yeah. Cool.

I think that's correct. Have you fired these up yet? Yeah. And you can I saw there's some issues with them. Okay. So I was trying to get him fired up last night before, you know, before we started the podcast, so I'll get all the way. And I could not talk to them. Who can I get the Raspberry Pi to talk to him? And these are using the PCM 5122 DAX and this is I squared C, I squared C and i square S. Okay, so I saw as I squared C for communication for like control, and then I to S is the protocol to do the sound. So the the problem is I just couldn't the driver on the Raspberry Pi couldn't

I think I see what Yeah, I think I see what you've done to the chips. Oh,

okay. That's actually not what the main problem Oh,

sorry. I just looking at the chips on it. See? Oh, that's so

funky. Yeah, yeah. The PCM 5122 is used on like the high high five berry DAC plus, which has built in drivers for the Raspberry Pi kernel. So all you have to do is to say I want to use this and it and connect the pins correctly and it should work. Well it wasn't. And I hooked up like I hooked up my digital logic analyzer. Then I was Getting this, this is after I did like all the software side like D message in like looking at like the boot log, and it's like it was trying to talk to it. And just like said, Can't talk to it and I give up. Yeah. And so I put my DLA on it digital logic analyzer and I was getting I squared C communication up to the chip to the chip. Yeah. And then it would fail. Okay. I'm like, okay, that's kind of weird. And then I put, I'm like, well, the delay doesn't always tell you what's truly going on, because it's a digital signal. So you're, you're seeing the ones and zeros. That's it. Right, right. You're not seeing like how much of a one it is, oh, I put the scope on it. And it was attenuated. Ah, and so it was only going up to like 1.8 volts. And that's not enough. That's enough to trigger the one on my DLA. But it wasn't enough to trigger the one on this guy. Is it like a 1.2 threshold or something like that on DLA? It was on this guy. It's like two point something volts. And so I'm like, Huh, I wonder if my one attenuating resistors is too high value. Put 30. I put 27 ohms. On there. Boom, fired right up.

He went from 1k to 27. Well,

the resistors are only there. I put them on there. Because this is supposed to go into pinball machine. Yes, there's gonna be a ribbon cable. And I wanted all the EMF to like, quench like I wanted, like just Yeah, stop it as hard as I can. So I'm like, one kg will probably work. Yeah, yeah. I got Yeah, yeah. It's just one kg was way too much. Okay. Yeah. So I put 27, which should still do a good enough job at killing the EMF off the line.

And it works. So whatever the input impedance of that pin was, it's not high.

Okay, or not high as as high as I thought it could

have been? Well, it was making a voltage divider. Yeah. With the with the one case,

it wasn't like a giga ohm. It's probably a couple 100k or something like that. Maybe not even maybe it's actually 10k, then yeah, it might it might be pretty damn low. Okay. Check that out. See what the impedance on that is.

So you were able to talk to it? Yes, I

got working. Well, that wasn't the only issue. So that got it so that I could talk to it. But then it would fail to driver initialization. And I started looking into how the Hi Fi berries not open source. And so I'll have like a clone of once anyway. So I'm like, Oh, just take the clone apart. So I D soldered it, Chip and just like that's kind of like routed the pins out. And the chip has to address lines that you can pull up or down that change the addresses of the IC. Yep, I had one messed up. Okay, so I lifted the leg and bent the leg over because they needed it was the pad was pulled down. So I lifted the leg and then the pin next to it was pulled high. So I just dragged the pin over and soldered.

I was looking at that because at first I was like, oh, Parker, here's your issue. You have a you have a solder short. All that way. Well, that's not a solder short. That's the pin

yanked. That's that's really cool. Yeah, so that works. So it works now. Okay, so pumps out audio. Yep. Nice. And so we're gonna do a, like, I want to try to doing some like, you know, pumping frequencies through it and capture on scope. All the different versions, see how much of it like if you put a single tone through how much distortion you get. And then I want to try doing a sweep, and then capture the sweep on the scope. And then maybe do like an FFT and see what I get.

Yeah, look at look at your harmonic content. See, see if you get any jazz above or below your fundament? Yeah, so I was gonna try

doing that. And then we're just gonna do a straight up, I'm going to hook them up into a box, and then have people around the office like, listen to him and see which one they think is the best one, huh, cool. Just do a straight blind test.

You know, I'm looking at the one that has the fancy schmancy film caps. Yeah, and some of the cats look weird. They look. They look like they have a bulbous

nose. That was by accident for me. Oh, it was Yeah.

Okay. I wasn't sure if that was like so that's like the I guess the coating on Yeah, I

actually coding.

I used to say I've never seen a film cap that well.

Yeah, they look great out of the reflow oven. And then I had to take that chip off.

Oh.

Cool, but it didn't work. It didn't affect the value either. Yeah, I actually checked it now it might have affected the frequency response.

I bet you just cooked the the varnish that's that's what it looked like to me. Yeah. So it almost looks like it actually looks like there's unmelted solder paste on the sun. But

hey, I didn't wasn't worried I think was just the varnish and it's it's steady state is still point one microfarad you know, it's probably fine.

Well, and the the purpose of these caps is not to actually filter right. I think I think that all of these cap values were selected such that it gives you the full bandwidth range.

Oh, the ones that look melted you're talking about those are straight up digital bypass caps that are for that chip.

Oh, so they actually have nothing to do with the signal path? Correct. Okay, well, they

probably absolutely do nothing. So

even if they weren't there, it still might work. Yes, sir. Okay, well, yeah, whatever.

Yeah, those were that's the most expensive board. Those point one microfarad caps cost more than the one audio caps that are in the audio circuit. Those are expensive caps. Yeah, like 12 Oh, 10 package or something like that. Yeah. Other film caps are not cheap,

especially SMT. Ones. Well, film caps through hole are cheap. But film SMTs are not correct.

Because that the handle to under 60 degrees Celsius three. Yeah, and actually, that's the thing, most do not have to check the data sheets for the reflow temps. Most of the ones I was finding that are the cheap ones can only do like lead free. I that lead for your lead?

Well, or the in a lot of cases, they're meant to be like specific caps that you put in one location on like, you choose film for its properties. Yeah. Either you're an audio Wenke or you need like specific pulse properties or really low ESR you can film in those situations. And so in whatever design you have, that even if it's SMD that cap might be soldered by hand. Yep, for whatever reason. So in a lot of cases they don't need to be that high of temperature Correct.

You said the specify its thermal junction temperature. Yeah, whatever that's called. Actually, I think it's what's called Yeah, but there's a specific time for like how long the iron can actually sit there for I

think it's usually like three to five seconds. Yeah, something like that. Which doesn't seem like much when you're soldering Nope.

Bless you if you started with just a you know, fire stick from Radio Shack.

Oh, like the 60 Watt. You know, Bazooka pin

no 20 Watt pin that you get from Radio Shack back in the day.

They were more than 20 Watts worth no

the cheap ones hell i my first soldering I was a 15 watt pin.

Wow. I remember they used to come with a tip that was so freakin huge that it was so hard to do a lot of hobbyist projects I first couple projects it did I was like man how do people solder and then I realized like you can get smaller tips Yeah. Also those those those old soldering irons they ate through tips I've found

yeah they're really crappy alloys yeah my my first let's just talk about solder knights i We already are Yeah, so my first thought on was a 15 watt soldering iron from Radio Shack and the other tip on that that's one thing I I just remembered is the tip was just horrendous on it. Like I would I would take sandpaper to just clean it up Oh yeah. And then you know of course that just made it worse because you exposed because it was like copper I think

oh yeah you can you can see like the plating and then the the underneath yeah

and so you just take down the copper and then basically you had to sand it every single time to take off the patina off the copper.

Oh yeah, I used to keep up I used to keep a file right next to my bed and I would just rub it on the file and it would solder like me like a champ for about a minute.

Yeah. So and then then I went to a 25 watt radio shack iron spent like the extra like five bucks once that 15 Watt one died. And then I went to a WP 30 which is a Weller iron Weller fire stick graded there that that iron last me a lot I think I went through like eight to 10 tips with that I still have it and that iron it's in my like toolbox I put in the back of the Jeep. Oh yeah cuz it's a it's just a really good iron in it. He ended up okay. Like heard the radio shack it like heated up

like I was like Oh, check you got to turn it on. Let it like bake for an hour.

Yeah, these you plugged in about 510 minutes it was hot enough. And then I went to I bought a Weller, W W E s 50 which is the bike basic model station. Now was one of the things were like this thing goes up to like 900 degrees Fahrenheit. And it like heats up almost instantly. And I actually still use that with like a I wrote on the side, like what tips the buy because like back then it was like really hard to find like the tips you wanted to buy like like Mouser or whatever. Yeah, so it's ETS style tips as well. I use Oh, that was your your specialty tip. Yeah. And and I built so many circuit boards for that thing. Like basically the beginning of high school to all the way through college. Until until I graduated yeah I used that iron well yeah now I use like crappy you like I like went like the other way like I was like getting smaller tips and then I bought a a what is that brand X Tronic Yordy 40 with the shovel Clow that comes with a lot of different tips and I was using that was like man these small tips suck and so then I put like this ginormous chisel tip that's like quarter inch wide really thick and butts comes down to a little tiny point and that is like the best thing ever. Like I soldered so many pin hex point four millimeter on those pic 32.4 millimeter pitch TKIP 140 fours with a club with a club and the thing is you just get it going and just like drag solder perfect every single time well drags Hunter is different than like sniper saw our Sniper saw because you turn it turn it 90 degrees and get this the point and just like sniper solder. The only problem with that tip is you can't get into places

knowing it oxidizes like oh yeah, it does.

Yeah, that's because the coating they use on those. Yeah, so the X Tronic uses like haco tips. If I just buy the Hakko ones now and they work a lot better by I get the big chisel because it's just like, as long as it comes down to little point and you have clearance. It works great.

If you don't have clearance, then you need metal stuff shitload of heat into a joint. That thing works. Oh, yeah, cuz I haven't maxed

out on temperature as well.

I'm using one of those Chinese you use or however you use Yoast or whatever. Yeah, I think I paid 100 bucks for it on Amazon. And man, those

things are awesome. Yeah, they're great. They're like, into the hot air station. They

are like as Chinesey them as you can get. Oh, just great. Yeah, it's the hot air station. And that's what my iron and it's awesome. There's a there's a new guy on we're not new guy. There's that there's a person on on Amazon that also sells replacement handles because those handles break all the damn time. Yeah, on those things. You can get a brand new handle and it comes with a whole pack of like 15 different 10 new tips. It's like seven bucks or something like that. And so the next day, we'll break on guarantee we'll break but every time it breaks, you get a shitload of tips. So it's like okay, great. Yeah, those things are awesome. I have done a ton of work on and it just it's a champ.

Yeah, cuz when you're doing like hobbyists, like Kabi soldering like we do on the side and stuff. It's like, those are the, in my opinion, the best rigs? Oh, for sure. They're the tips are haco compatible. So you can get nicer tips if you want. There's tons of parts. They they work great. And it has the hot air. Yep. You don't need to spend 300 bucks on a station.

Oh, no, no, no. One of those things. And they're getting cheaper. Yeah,

I mean, we buy. We have like the high end hawkers here on the floor. And we have thermal tronics. Right. Um, we talked about thermal tronics. Before I really liked the thermal tronics. Because the they use that curry heat frequency stuff super fast, super fast. And they never like if you're doing big ground planes are great, but it's like for hobbyists, you don't need that. You know what, that's one of those like, when it's like, oh, I want to spend $300 on a station. Okay, cool. Get a thermal Tronic. But

yeah, well, here's what I see is the difference between like professional soldering iron and like home gamer Sonic. comfortable grip. Oh, not even that actually. Here's the one difference that I see is like a professional soldering iron is designed to work 23 hours, 24 hours a day, you know, like when there's like shift changes, you know, might be turned off? Oh, yeah, it will work nonstop. It's the I would not trust my you you to go 24

It's the I think it's the in my opinion Site Reliability because I think my ex Tronic is I've had that thing for like five years now. And it's survived like, four moves. I've dropped it. And I haven't been replaced continue on operation. Yeah. And I've left it on over the weekend by accident before Oh, nothing bad happened. Yeah, just oxidized tip is is how comfortable that handle is. Because the thermal tronics feel really good in hand and it's really well isolated from the heat. Same thing was the high end tacos. Whereas the X Tronic like you leave it on for a couple hours. You can't even grab the handle because the handles hot. Not even mentioned the ergonomics of that hand.

I always thought the Radio Shack fire sticks got hot in the handle too. They did you had to hold them like back when you had the strain relief on the cable. Yes, yeah.

And that's actually one thing my WP 30 which is my Weller fire stick was like if you look at like the insulation around it was like inch in diameter. So like it was almost holding like, like a javelin on your iron on your wall.

If you think about it, you got to make 700 degrees go down to room temp 70 degrees in half an inch or not even half an inch. So I'm sure that's that's probably one of the harder aspects of making a soldering iron.

It makes sure you have enough isolation. thermal isolation. Yeah. Cool. So onto the RFO. Yeah. Oh, so you found this?

Yeah, I found a really cool chip this week. So I kind of I was having a conversation with a buddy. And we were talking about like, just random stuff. And one of the things we were talking about was a microphone idea. And I realized I never really looked into the microphone IC world. I have, but with only in like the pre amplifier world where it's just like, this is a pre amplifier that's designed to work with this particular type of microphone diaphragm. And so yeah, yeah, but but I never looked into like, the whole kit and caboodle package. And I found one, and it's super freakin cool. It's called the PCM 2912. A USB microphone, Chip. And this thing is literally like, about as all in one as you can.

This is like, a couple of weeks ago, we had the toaster chip. Yes, is the toaster chip of microphones and headsets?

What? Exactly, and the thing that's so crazy about this, okay, so get this. It has, it has a an analog to digital converter on this chip. It also has a digital to analog converter on the chip. And it has a basically a whole USB audio codec core. So it has one input where you, you basically just put your microphone, it has USB output, but it will also output audio from this chip. So let's say you wanted to make a whole podcasting microphone rig where you plug your headphones directly into this, and you can hear from it, and you can talk into it. This chip will handle all of that, including all the drivers to Windows, including all the communication to windows and everything. And it doesn't require off board components other than what you would normally I think it's like has like a crystal and yeah, it's got a

crystal and you need some passives and resistors and stuff,

right. So depending depending on what kind of microphone sensing element, actually

I printed off the optic layout guideline. Oh, yeah. So there's a problem with this layout guideline, though.

You know, I noticed that actually the other day, too. Well, let's let's see, what problem did you run into

USB input area?

Yeah, yeah.

Yeah, the d minus is actually the VB bus.

I said the other day, I was looking at this. I was like, man, if you hooked it up like this, then you'd have some

problem. I like. I'm like, I'm like, uh, why is the d minus? Which is the differential signal? Yeah, for USB. I don't know. It's running right into like, the anodes of the LEDs? Yeah. Like, why? Why are you running?

I thought I thought about that, too. Because I was like, Is this supposed to like light up? LEDs? Yeah, exactly. That's not supposed to carry a bunch of currents.

Yeah. What? So I went and looked and see what pin to which is where it's connected to like the V bus. Yeah, V bus?

Yeah. So ti kind of goofed a little bit on this diagram, I started actually looking at other chunks of the diagram and be like, Wow, did they goof on other things? And I don't think they did. I think there was I just

stopped there. I'm like, Well, they didn't get that. Right.

So yeah, like most of the passives that are on this thing are basically just like, support passives. Yeah. You know, just like bypass caps. Oh, yeah.

It's like how this this, this other, this? PCM 5122. Dac is Yep, it's all support. passives. And resist. Oh, I always say passes from resistors. Its capacitors and resistors. Or just passives?

Yeah, they're all passives. Yeah, they'll passives. So the thing is, like, if you wanted to build like, a headset that had a microphone on it, this chip could handle the whole thing.

No, I was actually thinking about that. And I might do a article series format crap on this chip. That'd be can basically prototype a headset? Yeah, do like 3d print stuff and build the circuit boards. And yeah,

that'd be super cool. Yeah. And the thing is, this chip in macro set is $11.50. But if you think about it, like for most microphone sensing elements, you don't need anything other than the element because it goes directly in and has a high enough impedance. Yeah. So you don't need any unless you want some you don't need any like analog conditioning or anything like that. On top of that, it has it has it actually has to gain amplifiers inside. So you can choose right at the front end, you can put 20 decibels of gain on your mic, right which is a 1000 times gain if you want, and then above and beyond that you can add anywhere from zero to 30. A little bit further on. So technically, you can go from anywhere from zero to 50 decibels of gain on this thing. So if you have a really, really sensitive element, yep, that needs 10,000 times gain, you can do it on this thing. And that's all programmable. You can do that directly from Windows basically. Hmm. And the zero to 20 is just if you pull a pin up or down. Gotcha. So I was looking at this kind of like, wow, it's expensive, but then it was like, it

literally does everything well thing. So it's kind of cool. You also paying for like the driver support and stuff like that. It

is okay, so there's two kind of downfall will mean, also on top of that, it you can run the entire system over USB power. Yeah, that's cool. But but to downfalls to it. And there's sort of downfalls not really depends on what you're looking for. First of all, it's only 16 bit resolution. So the standard nowadays for audio is 24. Bit, yep, whatever. 16 is still probably fine. Especially for, you know, if you're going to use this as like a gaming headset, you're going to be compressed to like three bits anyway, like, you know, so you don't need 24 bit audio. The other thing that kind of sucks. Not a big deal, though, is the maximum sampling frequency is 48k. Which that's more than enough. Yeah. If you're using Voice over IP, yes, way more enough, right. So you're not going to be doing like amazing recording on this thing. Yeah. But if you're using this as just like a capture element, you know, it meets Nyquist. It's good enough for 20 kilohertz, you know, so you wouldn't record a nice album on it. But hey,

you're looking for that Lo Fi effect. Actually, this is probably actually too high five for that,

well, you can turn down the sampling rate, the sampling rate to like, I think it goes down to like 11k, or something like that. So if you want it to be gritty, you could actually do that. Hey, yo, and all of those settings are actually done via windows. So you can change the frequency over USB

Oskol. It's partly through the windows sound options.

Yeah, that's exactly what Okay, and the thing that what's really nice about it is you you do zero programming, you just slap this on the board, and it works. That is nice. Yeah. So if you wanted to manufacture a bunch of like podcasting or streaming microphones, you just build it and ship it, you know,

build it and then ship it, right. Alright, so speaking of bad data sheets, or interesting data sheets, we've never talked about that. Yeah, so I was looking at because I was talking about building a USB hub for my Raspberry Pi compute module setup. And I'm going to use the what's called the The, it's microchips land 9514, which is a host controller. It's a host controller, but also has Ethernet built in. So you get an Ethernet controller, and you get four USB ports. It's pretty sweet. So I was going through the datasheet. And this is something that Stevens made fun of me a lot. And it says I can't remember what page it is on but I have a screenshot here is when does it says this in the when you're like doing like Power Monitoring for USB ports. It says when disabling port power, that drive will actively drive a zero not pull it down to ground or pull or just say pull down. It says drive a zero.

The thumb drive nothing. Yeah.

It's like cuz you may find me of like saying pulling it up to a one. Yeah. And I kind of said it as a joke. But this no, this is an in another engineer wrote this. Oh, yeah.

Another engineer. That's very like you pull it down to zero or drive it not Hi. Drive it to one

drive into one. Yeah. That's great.

There's there's one other out there somewhere in the world. Somewhere another one of you.

Oh, that's scary thought. Okay, it's a next topic is The Ben Heck Show. We've had Ben heckendorn on the podcast twice. I think. The Ben Heck Show who he is the YouTube video show that he hosts. He's not doing that anymore. Apparently.

When Ben's not doing it, yeah. Oh, he gave up. Well, it gave up

well, yeah. They came up with Yeah, they came out the video of like, the future of The Ben Heck Show and they basically was like an interview with him. And he was like, yeah, he's done.

Well, you cheers to Ben. Yeah, it's an entertaining show. Yeah,

I think it's like four or five years in the running so far.

Yeah. And they've done some really cool stuff on that. Yeah. You've been on the show, too.

Right? Yeah, actually, the I was on episode three, the only bill that failed and I did not have a crazy beard then. And And you were also the pinball Baron, right? Yeah, I was on the pinball episode too. Yeah. Yeah, that was a lot of fun. That was actually the that was the most fun.

That was. That was a long time ago. That was with Jerry Ellsworth. Yes. Yeah. Cool. That was

maybe that's longer than five years ago. That's before we started. That was what I saw on college. So man, that that pot that show must be going on for like eight years now. Yeah. End of an Era.

You also didn't you help with another? Will you help with I helped

with a lot of their projects on the back end, like conceptual wise, like, but

didn't you help with the one about like cleaning up bromide off of old Super Nintendo consoles or something like that? I

did that that wasn't with The Ben Heck Show though. Oh, I thought that that was something that was done the next show? I think they did. So we that was actually the Atari age form. Back when I was doing video mods for retirees. And I was helping out figuring out how to make that work with us. And yeah, that was that was for Atari like computers, and SNS cases. So that was that was way longer. That's another like lifetime.

So I got an SNS in 1996. Yep. And mine doesn't have that problem.

Yeah, it really dependent on what? They basically just couldn't get the mixture, right. Yeah. And so when it when I got it right, it looks great, because mine looks still brand new when I bought it. Oh, yeah. minesto has that like perfect gray and purple look. Yeah, yeah. But if they got it wrong, yellow after, like four or five years.

And what was the point of adding that stuff? Was it a fire returns? Oh, fire retardant. That's what it was. Yeah, yeah. Okay. So if you have a little too much, then yeah, UV light would cook it.

I think there's too little, maybe too much. I can't remember. It's been a long time one way or the other. Okay. Um, so yeah. Hopefully Ben goes on to different things that he wants to do. Is he busy? I think he's, like, tired of doing it.

It's it's been like a, regardless of if he's being filmed or not, he will do that kind of crap. That's true. Yeah. Like, he will absolutely have a good time doing crazy stuff.

So we'll have to have him back on the podcast. Yeah, for sure. Well, he's

already been on twice. So yeah, I'm sure it will be again. No.

So then the last topic is adventures in auto routing. Oh, yeah. So there was this article was written October last year. And it just popped up in the electronic subreddit today. It's I was looking at it, and I was reading it. And I'm like, this is definitely written this not offensive to this person at all. But it was written by a software person, because how is languages? The person I think his name's John Jones doesn't get it. So well. It's it's, I don't know who the author is. But it's josh.com. WP dot Josh calm? Well, clearly, it's Josh. So Josh. Is because all what the first Josh is writing about. It's not like he's talking about or just talking about like optimizing stuff. But like, like sup like imagined this is a quote, imagine today's standard four layer boards routinely being fit into two layers without any human effort. And it's like, most time when you choose four layers, it's not because it's easier to route. It's because you need noise suppression in EMF, or RF, or you want you need better power return and ground loop control.

Right? Right, right. Because of the cost increase, you have to like making the jump from two to four or even four to six. Like, it's a big jump, you have to be very deliberate about it. Yeah.

And, and so when he's talking about, like, all the things, or Josh is talking about all the things that he wants to do with this, he's talking, Josh talking about, like all these like, crazy machine learning and stuff like to make auto routing better, because basically, he's going through like four different auto routers that you can get, and it's and the deficiencies and he's like, Well, we could take all these open source designs and toss him to machine learning. Sure, that's often actually would probably make a really good auto router, but doesn't take in consideration the problems with auto routers. It's not that they don't always work with your your layouts, or they don't always route correctly. That's because they don't take into consideration they don't know what is traveling over those signals, right. And how the ground path return is affected by x action. Or oh, This is a 16 megahertz crystal. So we need to make sure nothing comes near that thing. And more than those traces need to be incredibly short. Yeah, really short and any trace near it need to be orthogonal to it, they don't pick up. What's that term

that was that was a, there was a fantastic word there orthogonal, orthogonal for their mutual conductance,

mutual conductance. That's what it is so. So Josh did bring up some good points there, but fails to really hit on why otter hours are not very good for most things. Now, I believe in the future, yes, we'll probably have auto routers that will have AI and can figure that stuff out. It could be even like, if the AI is or the auto router is smart enough, you could say it has a model, like almost a spice model of like, but SPICE models work really well, or pretty much analog ish style things. If you had a spice model, or equivalent of an 80 Mega 328 P. And then you can say, Here's my Arduino code, the hex file, and then you can actually run a digital simulation of your stuff, and then figure out I think, actually FPGAs do this that does already exist. Yeah, this exists. But it's, those are very expensive add on packages. And they're, they're less powerful than you think. Yeah. But given enough time, the human species can do anything.

So well. Okay, so So even the research I've done with auto routers, they're not terrible. But the you This is usually what happens is, an auto router will actually do almost as good of a job as a human. And I actually truly believe that. The thing is, you have to spend so long putting in all of the little like, intricate characteristics on every net, that it would it takes less time to just route the board yourself. Yeah, you know, when you're making your schematic, you have to say, Oh, this trace is special because of XYZ, and this pin is a output and this blah, blah, blah. Like, most of the time, you can hold that in your head well enough, even in a very complex design that, you know, the auto router is just not it's just not as efficient as people think it's it is absolutely not a button that you just press and poof, outcomes a PCB like, yes, a PCB will come out, and it is almost guaranteed to be garbage. So you have to spend so much time configuring it, that it just it loses its efficiency.

Yeah. So, Josh, he tried, he actually built his board that was based off this, and it works. Now, it worked in the sense of He loaded his code, and it worked. But he didn't do any of the testing on it. Like how much EMF was outputting? You know, it's a product that he's just building for himself?

Well, here's a great example. I know dip trace prefers by default, unless you configure it right, it will prefer that like certain traces go vertically on some layers and other traces go horizontally. So every trace is practically guaranteed to make a whole bunch of layer two layer transitions and 90 degree turns all over the place, you're guaranteed to make like just a noisy box at that point, you know, like nothing's making nice curves or, or 40, fives or anything like that. You're just making really sharp turns all over the place. And it's almost guaranteed to radiate all over the place.

Yeah, that'd be interesting thing is to do a test on like, let's say we took this DAC board, and then hook this thing up. And ran a we can auto route this one to rip of all the traces otter out and then compare it to what I routed here. Because I'll put it this way, this this board here. On the it's a two layer board. The backside is 100%, unbroken ground plane. Yeah, you did everything on top. Yeah. All the parts are on top, and then all the signals on top. Well, okay,

actually, I see a three part thing that would be fun. Okay.

So, but we need to be able to test that and we don't have the right stuff to test that thing. Well,

no, but but just consider this as as like a thought experiment here. Do three PCBs, the one you did already is complete. That's that. Do another one where you spend all the time and you make the auto router play by really, really nice rules. And then do another PCB where you just load up the PCB and click automatic. Just have it do like the shittiest thing possible

thing as I did, when I think the biggest thing with auto routing, and actually just routing in general is putting the parts when they need to go where the routes make sense. And so it's like I've already done that's actually the hardest thing. connecting everything together is not that hard.

Well, but but the thing is, it actually did the connections in your head when We're placing components. So

Josh actually brings that up, it says, Hey, if you have some components because his board design had LEDs on the outside, and he wanted only those LEDs to be, you know, spots, so all the other parts that matter where they were at. And so if the auto router had the capability of moving those around automagically to basically do what we do in our heads. Yeah, that would be pretty sweet.

I mean, we have like, really ridiculous iterative loops going through our heads while placing this, like, place this component, think about all the other components place the next component, think about all the other you know, like, it's constantly a matter of that. Yeah, no, it did. Trace has two routers, it has an auto placement and an auto router. So you can place all the components you want to be in very specific places and then lock them and then tell it to place all the other components and it will choose the best location for them.

Ah, maybe we should do this. Maybe we should convert this board to dip trace because Eagle doesn't have that. And you know, just the auto router doesn't have the auto place. You know what's what also Eagle just put in their system? What auto BGA fanout Okay, that's

useful, but haven't tried it yet. That's really useful. Because you can spend a whole freakin day doing that by hand. Yeah, that's that's kind of

nice. I haven't tried it yet. To see.

I guess it has just a whole bunch of parameters you put in and automatically no idea. Oh, give it a try. That sounds cool. All right. It's not like you and I use BGA is very

that's the thing is usually I stopped that TQM Fp 44 141 40 fours. DQ FPS which are point four millimeter pitch parts,

right? Because they're hard to do, but you can't pin pins.

I think these are point five.

Yeah, I could see it with my eyes. These are point five.

Look at it. And like It's like Terminator vision, like auto calipers come in your peripherals.

Except the boards are red. So having Terminator vision would just screw everything up. Yeah.

I can't see it anymore. You know what the funny thing about that is? This podcast has gone on for like 58 minutes. But one more tangent on that. We're gonna make it a full hour. Yeah. Is Terminator vision. Okay. And so we have a bunch of scanners on our SMT line. Yes, I'm Steven Whiteside about this. And they weren't great until they don't work great. Like, they'll just miss boards that go underneath them randomly. And so I'm like, Okay, why is it happening? So these scanners have three modes they go into, they have a deep sleep mode, they have a active scanning, so they're actually looking for barcodes. And then they have a passive scanning and passive scanning mode, do not actively looking for a barcode, they're just looking for a contrast difference to then be activated into barcodes and they jumped into app. Yeah, and that's power saving or some, you know, whatever the Edit, whatever for that reason. I actually think it's for heat because they are in like active scanning for a long time it says that, like they can overheat. So it probably something like that. Anyways, they flooded the area with red lights, and we can turn off sleep mode. So sleep modes disabled. So it's always in passive scanning. Well, when you have a red PCB going into an area that's saturated by Terminator read vision, light, it doesn't actually see anything, like I actually basically reconfigured them, because they're just image sensor. So I had like, piping in video. And like, when you see the board, I can see it because I know what to look for. But it's just, it's such a faint line moving across the scanner sometimes picks it up and sometimes doesn't. So the solution is the mega hit all the time was to, um, disabling the red illumination and putting like, low LED lights

and actually just break them open and swap the LEDs. I actually thought about I thought about like blue light or something like that white.

Well, that would work. Unless there's a white PCB. Yeah. Oh, no. Then you just put a black substrate down. And then the white comes. Yeah, that will don't work. There you go. This is a blackboard can't win.

multicolor LEDs that just like cycle?

Yeah. Ah, god. Yeah. Yeah. On a very quick, yeah, rotation. Yeah. But um, because of what I noticed in the data was the boards that were not read 100% And like, the red boards were like, 80 some odd percent.

Yeah. Interesting. There you go. Yeah, Terminator vision is not perfect. No, no, it's not. I need you to close your boots and your motorcycle,

and your PCBs. Oh, wait. I can't see.

Okay, so yeah with that Terminator. Yeah. 107 terminate the episode that we were your host, Stephen Craig. Take it easy.

I'll be back Thank you. Yes, you are a listener for downloading our show. If you have a cool idea, or project or topic that you want Steven and I to discuss, tweet us at McWrap or email us at podcast at my cred.com Also check out our Slack channel. If you're not subscribed to that podcast yet, click that subscribe button. That way you get the latest episode right when it releases and please review us wherever you listen. Whoa, that changed podcast addict or iTunes. It helps the show stay visible and helps new listeners find us. Matt crib is also hiring apparently still