How Soft Are Your Diodes?

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

And Steven Craig.

This is episode 249. So this weekend coming up November 7, Saturday, all day Saturday, well, time I wake up and when I fall asleep like 24 hours later, I signed up for the extra life charity, which is a charity stream, video games charity stream for children's hospitals all over America. And I'm raising money for the local chapter here, which is the Texas Children's Hospitals. So there's a link in the description of this podcast and on my Twitter that you can go and donate. Basically, it's like, you know how you have all those charity like walks and jogs. I think there's like one here in Houston is called boot walk for cancer. You know, you basically go around and get funding from people to donate to you and you go walk while instead of walking, I get to play video games for 24 hours. Something that I prefer better than walking for a long time. But yeah, all donations are tax deductible. 100% of it goes to the hospital. There's no middleman or anything like that. It all goes to the hospital.

So what do you what are you going to be playing Parker?

I'm going to be playing an older game. It might some this game might have come up before some of our listeners. Oh, gosh. It's kind of crazy to think about that way. But yeah. Oh,

yeah. Yeah, if, if you are 15 or younger than, then

this game is older than you. So I'll be playing the half life two series, and essay series, because there's the base game Half Life two, and then there's an episode one and episode two. I'm gonna play through all three of those. And also on the hardest difficulty, and 100% achievement locks. So basically doing everything and finding all the easter eggs and yeah, all that stuff. Do you have that

memorized? Like no it off the top of your head, like all the unlocks?

Ah, I want to have a list of all the achievements. So I make sure to have the remind myself of them. And actually, that's one thing I'm working on. I spent Wait, I was up way too late last night working on some Java scripts to try to figure out how to read the achievements out of steam. So steam is like this ginormous store to buy online store to buy video games from. But they also have the achievements are in there as well, which are like, milestones in the game, right. And so steam has an API that you can pull and get the achievements that you currently have or don't have for the game. And so what I was going I've been working on is trying to write some javascript so I can pull that list. And so that on my video game stream, I could have the list there would like checkboxes next to like the ones I've done so far in the game so far. Nice. Sounds like a lot of work it is because no one's actually tried this before, like I've done so much. I spent like two hours Googling, like last night, trying to see if anyone's ever tried to do this and no one has. And how steams API works. You can't just do like a fetch quest, like a fetch of the the API endpoint. You can't do that you have to like actually run like a JavaScript server and stuff to make it work. So I got to run like a local server and make it work. I think I'm going to be able to make it work. Hopefully, by the end of tonight, after this podcast, I have like a prototype that I'm like, okay, I can do this. Because it would be nice, because then people could come in and see what I'm doing and how far through the game with the achievements and stuff like that.

She you should have it were tweets out an achievement every time you get it.

That's good idea to I mean, it is I'm learning because I don't know enough JavaScript to like build local apps and stuff but I've never tried using what is it? Like no, Jas never tried using the me that kind of stuff. So I'm learning all that tonight.

Like that. So it's a window when does it all start?

So technically, if you go with extra life is like the big umbrella of like the whole thing. They kind of like organize it and make sure all the infrastructure is in place. And like provide the way for people to funnel money into the charities, that kind of stuff. So you can do it anytime. But The Big Event is November 7, which is a Saturday and I'm going to start at 8am and go to eight p 8am. On Sunday, Cheez Its 24 hours, I think. So the last time I play I did this I could do. Basically the three games, I could beat them in 16 hours. Oh, God hard. Now, I don't know about all the achievements, also. And also, I want to try to do without like safe scrubbing. So like, I don't want to be like, Oh, I messed up, I need go back basically might mess up. It's like, okay, we just won't get that one. Right. But we're gonna try to get them all.

Yeah, I was wondering because a lot of times achievements are like, make sure you blow this one things head off. And if you don't do that, it's like, it's not as simple as like, go here and grab this item that's on the ground, you know?

Yeah. So the big one is going to be like this one, we're fine. All the lambda symbols, which will like I have to remember where all those are at. It has been five years at this point since I played this game. So yeah, it's gonna be it's gonna be fun. Great. That's cool.

Yeah, well, once again, why don't you tell people where they can find your stream at?

So there's definitely gonna be a link in the description of the podcast. It's gonna be on my Twitter, which is Longhorn engineer. With no O's in Longhorn, it's kind of confusing. But yeah, I'll be tweeting it out. It's also in our Slack channel. koolatron. Like, donated, like, oodles of money last night, actually, right after I posted it. Like he actually made it. So I hit my first goal. Like, right away. handedly. So I had Yeah, so I had already I already moved the goalposts down the line. That's, that's awesome. Yeah, it was,

it was awesome. So is it like Twitch where you get notified when somebody gives a donation?

Yes. Yeah. So if my twitch channel is twitch.tv/crab foam, not to actually grab foam. Yes. Bill to actually craft from CR ABF. Oh, am and that will also be there. So I'm hoping people stopped by on Saturday. And like, watch, hang out, chat with me. That kind of stuff. Watch me, hopefully not die. A lot of times. There's some sections where like, I will die repeatedly cuz I'm trying to do a certain like trying to get a certain achievement.

And heart is moderately hard and HalfLife to Yes. Does not allowed bullets you can take No, for sure. Not. Very cool. Well tune in for that and give some money to charity.

Well, okay. Hardware now. So this weekend Pignataro. Revision three went gold.

Yes, like production version.

It could be depending on how well it prototypes. I ordered some prototypes for it. But it's gotten gold boards are, the designs are off into the into the fab house cloud, right. So some differences we did overdrive to is we added more current sensing to the board, because previously we were only current sensing on the solenoids. And they wanted to do current sensing on like the motor controls and the servos. Just to see like that. So you can they can actually detect if a servo is stalling out or something like that. But we had to actually change a different chip. Because previously we were using the Allegro. And this is a doozy of a part number, a c s 711 KEX lt hyphen 15 AB hyphen t, which is the 15 amp version of a of a, I think it's a hall effect sensor, current sensor. Think it's Hall effects how they work, I can't remember. Basically, it, you put a bug, you put the current trace into it, the current trace comes out of it. And a analog voltage signal comes out of it that you can read off your microcontroller. They're pretty sweet. But the 15 amp versions were too beefy, for like the low current servos and steppers. And so we are using the five volt version and a five volt five amp version of that same chip. It's like it's like the same footprint but has some less features. It's kind of weird. And this is even longer part number I'm just gonna put in the part notes or

the the another ACS seven. Yeah,

but it's like completely different format. It's really weird. But the difference between them is not as just it's not just the current capabilities of them. The high current one also has like a fault line, like a fault signal. whereas low current ones don't, I don't know why

it may be some kind of requirement out there for like, a lot of times these are used in automotive or some other industry that that requires that extra line.

Yeah, because the package has it for the low current, it's just not connected and seed. But because of that fault line, we're actually going to use the fault line on the 15 amp one now. Previously, the microcontroller controlled the relay that loud all the current to come on to the 50 volt line for the solenoids. And but we'd never ran into this issue, but it was just one of those, let's just add more safety into this device. Like, what happens if the microcontroller did lock up, and the output was on high, and then a solenoid decided to run away and start smoking, right. And so we added a fault pin, we've now added the fault, we took the fault pin output of the current sensor and ran that into an AND gate that went into the relay enable MOSFET driver now Oh, so we can kill itself if needed be Yeah, so the microcontroller has to send a one to the end gate. And the fault has to be because the faults active low. So it has to send the one basic goods pulled up. And if those are both one, then the gate driver for the MOSFET turns on. And it turns on our brand new different style of relay that we put on the board too. Because before we were using like these other just standard standard relays, but these were activated by three volts. But we were actually having issues with like the contacts were like welding together, it's kind of weird. We could never figure out why because it technically was rated high enough. But it was still, I don't know, I think we're just like switching at weird moments, or maybe the coil because it was such a low voltage coil. It couldn't switch fast enough. And so we, I took that style relay out, and I put in a socket style relay. But for automotive applications, so we actually using ISO micro relays that are in automotive. And I've used these for like, switching 50 amp fans. In automotive applications. I'm like, Okay, there's no way we can kill this thing. But um, that was also the other changes, we had to change how much what voltage we're driving our relay with, because those are 12 volt relays. You can't get an automotive three volt one, right? Right, right, right, you have to change what you're getting your voltage from. And the last major change is we put the flyback diodes for the coils on the PCB, instead of having the coils have them on the back of them. And this is mainly for, in case someone is building their PC, their board are their pinball machine and they accidentally wired diet up backwards on the coil, which happens just to prevent the board from blowing up.

Yeah, honestly, I think it probably not be a bad idea to do both. Tell them to get them on there. And then there's protection in both cases. Yeah,

the main thing is, this will least prevent the board from blowing up but they put the diode on backwards on the coil. But putting the dial on the coil is technically the best thing to do for EMF, because you're putting the diode as close to the EMF source as possible, which is a ginormous copper coil. Right,

right. Right, right. Yeah. But yeah, two levels of protections. Not not a bad thing for sure. Yeah. And putting the extra diodes on the board is going to raise the cost by nothing.

It's I think 20 cents overall for the 24 diodes plus assembly.

Sure. Yeah. So it's just worth it. Right. Yeah. So it's worth it.

And I didn't actually have to do a lot of layout changes for that one because I put them on the backside of the board, just plenty. So I just put them right next to the pins as well. So they're in the optimal area for that to work.

And those who aren't on terminal blocks right. They have like keyed connectors,

their key connectors Yeah. They're Molex KK types was point 156 inch pitch.

pretty beefy connectors for pinball point 156

Yeah, that's a standard pitch for it. That's weird. It's their way overkill, though. This is your perfect kill protectors. Yeah, their monster. Yeah. This is the this little pin hack but oh, yeah, these big guys right here.

Yeah, they're really beefy. Yeah, but standard fuzzies though when you're putting it together, it's like big chunk of chunks, kind of actors.

Yeah, so I'll take a picture of the connectors for the For the blog and stuff, but I'll take a picture next to like a normal 100 mil one, which is what the servos use servers use 100 mil connectors. But the penetrator actually since we shrunk the board so much, we actually the only thing using the big style connectors are the high current stuff like the coils, everything else is using 100 mil Now,

sure, they'll change the standard,

slowly but surely,

honestly, to change the standard like that, somebody just has to do it.

That's actually the thing is just like, because most pinball uses glass style, like 20 millimeter by five millimeter fuses, which are a pain in the butt to get locally. Because who sells those, you know, like RadioShack used to it is

a pain in the butt most of the time. There's there's two electronic stores in Houston that carried them for sure. And I know I can always get them at those too. But I'm in Denver, and I can't think of a place right now that I could get them.

Yeah, they're a pain in the butt to get. But they're a standard in quotes standard in pinball because some engineered decided back, you know, in the 70s start using them, you know, to go with the smaller ones. Yeah, yeah, well, so Penetang this is actually the first revision, we changed the fuses as well. So instead of the glass type, we're actually using automotive mini blade fuses, which are amazing, because they're dirt cheap, because they're like, five cents apiece. And you can buy them at any other store. Right? Or like, you can buy them at Walmart, you can buy my target, you can, you can go to a gas station and buy them. So it's like, okay, you can get them anywhere.

Right. Right. Well, and it works for pinball because the values are higher, right? Because you're not going to find a quarter amp on those right? No, like readily available. But

our minute our smallest is a is a two, no three amp fuse, I think is the smallest on the board, which you can get those you can actually get down to you can get half amp, and then the blade fuses. So I'm excited. That's

one of the one of the boards come in two, three weeks, something like that.

Yeah, November 20, is when they will be in my hands.

Okay, so maybe we'll hear something in December on that.

Hopefully on the 24th. Cool, well, everyone will hear about on the 25th. Actually. That's November 25. That's a Wednesday.

Yeah, you have a testbed made for this version? Or are you just going to scope it out and still

just scope it out? Yeah, I'm not too worried about it. The when we go production will have a bed of nails or something like that for actually those with the, with how I've designed it. There's actually no reason to do a bed of nails. It's basically just power it up and make sure you have current limiting on it's pretty much only if it programs, it's like okay, there's nothing else really for it to go wrong. Because if it passes automate aeoi It's there's nothing else to test because everything's you know, lead ID. It's not like the pin hack, where like, the pitch was really tiny on a lot of parts. There was one major thing I did on this design was like use big, wide pitch parts that you can see in SPECT to prevent us from having a an expensive test fixture or test process for him. Yeah. I wanted to take him off the line power up good dome in the box.

Yeah, yeah, I suppose I suppose if you've prototyped them, and you've proven that, you know, everything's good on the prototypes, then. You know, as long as that gives you enough warm and fuzzy, then you're good to go.

Yeah. Well, it's also the there's nothing to be calibrated on them. Right? Because everything is just on off. Yeah, it's digital man.

That's cool. So actually, what the current limiting is that? Is that something that you are that you handle inside internally, yourself? Or are you giving that capability to the game designer, they can constantly sense current sense and like shut off and things are?

Yeah, so the, the game designer has a little bit of control, like they can say this should be these parameters. Okay. And then it handles that. Like all the low level stuff, handles itself.

Is there some kind of like higher level code that you're offering or do the game designers have to design on bare metal

so So we provide peanuts are actually you won't be programming peanuts or you actually, we're working with a software package called Mission pinball framework. And you write that and that runs on like a Raspberry Pi or computer. And then the penetrator is a USB device. And so it gets commands of like flip this switch or, or this solenoid. And then it gets a, you know, it gets a switch report back from Penetang. And saying, switch x was pressed. Do this now. Got it?

Okay. Yeah. Yeah, runs off of a computer. But you could if you wanted to program a game in bare metal,

you could Yeah, the problem with panther is it doesn't have any ability for sound, or display got natively. So you would have to hack that in somehow or just not have that. Like, you could run a buzzer off a song off of MOSFET. But you couldn't do like music

PWM music

now, you can do that. It's just how far do you want to go? Right, right,

right. Yeah, yeah. Okay. So you're the, you're the muscle in the brawn controller, and the computer just commands the muscles.

Yes. Cool. And it allows you to do like crazy graphics and stuff like that

that way. Right? Because they had the computer can handle that a lot easier.

Computer easily.

Yeah. Yeah, I suppose you could probably actually do a Raspberry Pi and get pretty complex with stuff.

Yeah, the Raspberry Pi 4x. Runs mission pennwell framework pretty well. Does it from what? Yeah, from what we've experimented with.

That's cool. Yeah. All right. Yeah. Well, I'm looking forward to that.

I can't wait. It's been. We did Brev. Two in I designed rev two in your basement. Was Rev. Rev. One was in your basement. That was right. Right.

hoof. COVID makes everything weird. Everything is like feel slower, and way faster at the same time.

Exactly. That was two July's ago. 2019 July.

Yeah, sure. Crazy. So I ran into something interesting today that I had never seen before. And so I wrote up a quick thing, how softer your diodes, say clickbait title, if I've ever written one up, so I was get this. I was looking at MOSFETs today because I'm doing a design at work that needs some power MOSFETs. And I was just going through the datasheet. And I ran into a section, you know, where it calls out all the all the characteristics, and it had s factor in it. And next to it, it said softness factor. And I've never seen softness factor. So so before you read the notes, Parker, take a guess on what you think softness factor is. How softer

diodes. So it's a diode. And it's just a it's just a standard style diode. Right? Well, well, this

was on a MOSFET a datasheet. But this applies to diodes also.

Oh, yeah. It's a, it's a number that explains the curve the cut off.

Pretty, pretty close. Actually. It was okay. So what softness factor is, it's a definition of the reversed current characteristics on a diet. So get this. So when a diode turns on, it builds up a space charge inside the pn junction, right? That's hunky dory. That's all great whenever the diode is is on, but as soon as it flips to turning off, so once you get past that threshold, current still flows through the diode until you are effectively reverse bias, right. So then the diode, the, the problem is you still have charge left in that space charge. And that has to deplete. So charge will actually reverse and you get reverse current through the diode. That's why like if you ever look at a rectifier on a scope, you see your lobe that you're expecting and then it goes the opposite direction. And it seems really confusing, right? Well, that's actually just the depletion of the space current, the space charge current in the in the diode and the softness factor is actually a definition of how that actually looks. That portion and so I'm plenty aware of the reverse characteristics of diodes because I have dealt with them a lot. But I never knew that it could actually be defined. And, and funnily enough, it's actually defined as two separate sections, there's a TA and TB. And that's time of intersection A and A time of a Section B. And I found, there's this great article called, it's from electrical for you.com, which, by the way, I've never heard of this website, and I went to the website, and it feels very, it feels very similar to the idea we had of so you want to design a blah, blah, blah, it feels like structure for that. Not it's not organized in the in a way that was in my mind, but like, this whole article was like, here's softness factor. And here's all the things you need to know about it. Here's the things that matter. And I was like, this is a great article,

and actually, like, how does it matter to what you're designing?

Exactly. And there's like charts in there that it's like, if you if there's this diode, it's this kind of thing, and they're there, these ranges and blah, blah, blah. So yeah, go check this out. We'll post the link up in the show notes. But the the kind of the softness factor curve is defined in this article of that TA and TB section where TA is the time when the charge from the depletion region is removed and then T B is the return to zero current, which is the time when the charge from the semiconductor region is removed. And the softness factor is a ratio of those two times. So if the ratio is unity, the device is considered a soft recovery device. So it has a has a smoother curve and it's slower, shall we say? And then if if the S Factor is less than unity, it's considered a fast or a snappy recovery device.

Ah, so this is this is how you can get a Schottky diode. This is what makes a Schottky diode, a Schottky diode or a fast recovery diode or fast recovery diode.

Yeah, this this will in relation to the reverse recovery time. Yeah, this is yeah, this is exactly what makes it so the GAO general purpose, your Joe Schmo Jelly Bean diodes typically fit more into the soft recovery zone. So they have more of a unity S Factor. And then your fast recovery or your Fred diodes, which are the fast recovery epitaxial diodes. They have a lower than unity S Factor. Those, look them up Fred diodes are a real No,

no, I'm shaking my head because I would totally have come up with that as what a champion if it was my idea,

they're great, they're great. And actually, it's good. It's kind of interesting, like, you know, like the reg, I'm sure at most everyone is aware of the one end four series of diodes, the four zeros or 1234, you can also get them in a UF version, the ultra fast 4001 or 4003, or whatever. And the difference between those is S Factor and how fast it recovers. And what's interesting is you can kind of fine tune your circuit, especially if you're putting them in like a rectifier application, you can fine tune your circuit based off of the ringing of the S Factor and what happens in your rectifiers based on what the highest frequency you can accept in your circuit. So if you typically soft recovery diodes are a little bit cheaper, and they're slower. So if your circuit can handle that ringing at lower frequencies, then go ahead and use your soft recoveries, they'll be fine as long as your circuits fine with that. But if that's a problem, use your fast, snappy ones. And in fact, in in my guitar amps, I use fast snappy recovery because if that's too slow, it manifests itself as ringing and buzzing in the in the sound. So I use the fast stuff, because it usually ends up being higher than audible. And so even if it exists in my rectifier, then who cares? You know, you can't hear it. And with guitar amps, most of the time their band limited to like five kilohertz. So like yeah, you know you can get away with with slower stuff, but we're using

one and four Oh fours for our our flat backs. Yeah. So I wonder if that's big enough. I guess that's what they use an industry.

Yeah, definitely. All of those is the voltage range, right? Because one end for zeros or ones like 50 volts, and the 400 sevens are like 1000 volts. Yeah, and then it increases in between those.

We'll be using the are a 404 which is the surface mount version and service mount version on semi. Yeah.

Gotta get that Fred MRA 4004 Although for your application don't because it doesn't matter, right? Like,

I don't know what's what is like the ideal diode to SLUB a the flyback from a coil,

one that can handle the energy, like, ideal one, right? As long as it's fast enough that you don't get a massive voltage increase and it can handle the repetitive the total amount of energy that the coil is going to dump through it. Yeah. So the those, those 4000 series diodes are beefy as hell. Like they just take abuse. So you're probably fine with that.

We're just using industry standard. So yeah, so it's

been you've had plenty of proof that it works, right? Yep.

Well, they do fail every so often. And no one knows seems to know why either.

Okay, so put some 4000 and fives in there.

Actually, it seems to be like when a solenoid starts to wear is when it could go out. So I don't know what that is about. I don't know enough about electromechanical stuff. What does it mean when a solenoid wears? Like, like, usually it's like the liner inside of it. That the nylon liner starts to get a little grimy in there from repetitive motion of the chromed actuator basically it will start to gum up or, or the coils are being run above its duty cycle. It starts to heat up.

Oh, yeah, sure. Yeah. I guess if the if the balls just like hammer in whatever mech that is. Yep.

Cool, huh. Okay. So this is a funny funny in quotes article that I found on on fierce it's a Walmart fires store robots. Yeah. Because, you know, we were in a climate of, of, you know, a lot of people are on on unemployment and that kind of stuff. And, man, these robots can't get a break either, because they're, they're unemployed now to. Basically these are, these are robots that Walmart was using to control, like, shelving stock, like, like the, what was on their shelves, and make sure that the shelves were full. Okay. And they got rid of them all, because, well, one, they did see improvements in inventory control of the robots, but not enough of improvement in revenue, or other measures this put other measures in there too. Because the idea was, as you keep the stock, the shelves stocked, you will sell more product, which makes sense. But apparently, it wasn't and that that premise is not enough to keep the expensive robots around, at least more so than hiring, you know, a high school student at $8 an hour to walk down the shelves and make sure stuff is docked.

Actually, having worked at Walmart in the past, one of the biggest things that they would really wrap on us for is not having the inventory at the front of the shelf. Because yeah, there's this perception that if you see something empty, that's bad for the for the customer store. And so you always try to fill in gaps and you try to make things look clean and neat. I agree. Because like if you've ever been to a store where you know, the the shelves are just empty, it's like, oh, God, this is the Wild West. I don't want to shop here.

And Fry's Electronics. No, no, that's actually exactly what they were talking about. But they did see improvements in that. But it did not turn into enough revenue extra revenue to pay for the robots. Yeah. So they probably did see some improvement there. But not enough to keep the robots around. So they they fired all the robots. See if the robots actually had a I could they file for unemployment?

Yeah, and they're expedient with every document like it's there on time, like right away instantaneously. Yeah, yeah.

So I just thought was interesting that it was mainly the title fires store robots.

These robots look kind of creepy. They actually look like something that would be in a Doctor Who episode.

Yeah, they really just actually look like trash cans.

Exactly. Something from a doctor. We talked uh, gosh, remember this? We talked about some robots. I think they were security bots. And one of them like fell in a fountain or something. Looks Similar to.

Yeah. Yeah, basically they they said that, like, because the robots would go around and basically notify human that they needed to adjust the inventory, like the bring the box for it, right. But if there was a problem with the robot, you had to get a more expensive technician out to fix the robot. And it's like, well, you could have had your high school $8 An hour student just, you know, or employee, just walk down the hall, you know, the aisle and just do it no matter what.

Yeah, so Walmart, Walmart has the kind of money to experiment experiment. Yeah, with this kind of stuff. Yeah, we are certainly not there yet. No, but eventually, for sure. All right. So something happened earlier today, that really kind of sparked my interest. We're on episode 249. Now, so we've gone quite a long time. And in our Slack channel, at Zayn said that they were going back and listening to Episode One. Now, I don't know if that means that they had gone through all of our episodes. And then and then gone back, but but they're going back and listening to Episode One, which bless your heart. And thank you very much. And also good luck. Because those were those were the dark times. But

they were not dark. We just, we still want to know what we're doing. No,

no, no, no, we're only slightly better now. But no, no, no, thank you for going back. And thank you for listening to everything and going back. That's fantastic. But what this really brought up was a handful of comments about the SSPs. And I realized that we hadn't talked about the SSPs in a while. And the SSP. S is a dark project, I guess we could say, because it's it's the project that just won't finish and just won't get done as are like most of our projects, but it was from day one, our project. And I figured because we hadn't talked about it in a while, we might have some new listeners that have no idea what the SSPs is.

And I'd like the new name for it. Well, there

was like, a handful of new names. Which Which one was your favorite? Oh,

the spooky simple power supply?

Yeah, it is spooky now? Yeah, it's a ghost almost. So I started thinking about it. And, and what what I wanted to just cover real quick is what is the SSPs? To bring people up to speed who might not know what it is. And then I started thinking, what were its specifications? So I went back and looked at our notes from the first handful. I couldn't find any of our specifications for it. I don't know if we actually wrote them down, because we just knew them at the time. So I wanted to talk about the specifications. And then where did we leave off with it because Parker mentioned in the Slack channel, that it's just chilling in my basement, which is partially true. It's actually just chilling in my garage right now.

I was one room away from it one room away and

exposed to the elements. So okay, first of all, let's let's talk about the specifications for the SSPs. Because there might be something cool that we could potentially do with it in the future. So if I remember right, that's the SSPs. So first of all, let me step back for a second. The SSPs was the first project that we came up with on the podcast. And it was basically a monstrous linear power supply. That was digitally controlled. It had two outputs. It was capable of 10 amps, I believe was the number that we came up with. We were originally

found. I found the specifications, at least what we wrote down in what Google Docs or blog to isolated minus 30 volt two plus 30 volt at 10 amp outputs,

right okay, yeah, full 70 volts,

and then we're 10 millivolt per bit control.

Sorry, I'm laughing That's pretty. That's pretty fine resolution for that. But okay, okay. So here's the thing. I went back because I had very little recollection of what what we had actually done, and we actually documented a whole lot of stuff. We have a full GitHub, which we'll post the links for that. I have them up here. We have a full GitHub, I went back and started looking at stuff. Both Parker and I design boards for this I designed what we had originally called the Energon cube, which the Energon cube was eight monster screw terminal cup. acids that had copper busbars screwed into them.

It might be the scariest thing we've ever put together on a bed. Oh, it's sort of a boatload

of it was it was, it was the kind of device that when you turned it on the lights dimmed. So yeah, the Energon cube was eight giant reservoir capacitors, that our PCB bolted to those. And my analog board was a big analog basic regulator, effectively, that these, the idea was to have a giant copper waterblock going down the middle of it, that we could bolt all of our power past transistors to it. And actually, if you go listen to the first few episodes of the podcast, we talked about running this thing off of some pretty crazy op amps that were available, we actually ditched that and just went to a discrete version of things. And I went back and looked at the design. And it was I actually did that because I'm doing something similar, but way, way, way less power at work right now. But it's still similar layout. It's still, you know, past transistors with, with error correction, feedback and stuff. And so we've actually designed both of these boards than analog board and a digital brain to control it. And we hooked it up and the thing actually worked, like we had the SSPs functioning, and then we put it down because it just we'd never really got to I think the part that we left off what was the faceplate, we never actually made the faceplate for it. Correct. Is that really all that

we had? was when we had to do another spin on the the boards? Yeah. And we never actually made the heatsinking. Correct. Never made a heatsinking. Right. But we have all this stuff to do the heat sinking? I think,

you know, I saw I went out to the garage, and I checked the parts that I have. I have everything except for the radiator, which whatever the radiator is something we can't get off of Amazon. So yeah, we wouldn't need to redo the boards. There was there was one thing about the design that kind of sucked, I guess you could say the because it was a linear power supply. One of the downfalls of a linear power supplies, the way you get regulated power is everything you don't want, you just get rid of it as heat. That's the way a linear generally works. And so this power supply would work really great if you wanted 30 volts 10 amps, like it's great and probably run fairly cool doing that, right? Because it's not getting rid of anything. But let's say you wanted three volts at 10 amps. Well, it has to get rid of 27 voltage. And it has to get rid of that as heat. So it's massively inefficient. It is awful, inefficient. But it has the benefit of being a linear power supply, which if if you're into it, linear power supplies are typically superior in terms of their noise performance. They're generally better. So I guess we were kind of decided to go with that as a primary factor. So I don't know. I thought it was fun to go back and and look at our designs.

I think I think the reason why I liked the idea was was actually building a linear power supply that was like this. Just rent that was just super overkill. Yeah. Yeah. Like you can do 10 amps and a linear power supply. Or you can do like, you know, a quarter an amp.

Right. Yeah. Yeah. And you could do 10 millivolt. Resolution. Yeah. It probably was picked. We never actually tested the resolution on it.

No, I think that was just because that was what the analog to digital or digital analog converters were

capable back that right. Yeah.

So probably be somewhere higher than that.

So you know it. Yeah, it's likely higher than that. I but I don't know exactly what its absolute accuracy was. I remember originally, we were talking about making a linear power supply that was capable of producing a full 170 volt swing at at 60 hertz. So we could simulate mains into a device that no,

that was actually one idea we had is, technically you could do it. This thing could technically do it.

No, it doesn't. It doesn't have an A voltage showing on

it. Yeah, no, but you would have to basically double the insights. Oh, yeah. I

mean, we'd have to go ridiculous. Yeah. No, we'd have to, we'd have to triple insights. Because yes, yeah. Because Because your absolute value have to go up to about 170 volts. So yeah, that would be really cool to have something that could produce a perfect mains sine wave. Just generated from a box. That was the original thought, but that was just a little too ridiculous. Yeah, yeah. One day we'll do something with it. And I would love to see it evolve. You know, I would love to see it evolve, frankly, what would be really cool is if the if the Slack channel was into it if we all designed the SSPs together and really like refined things based off of the hive mind knowledge and come up with a perfect SSPs that that still like last week that still maintain the spirit of the SSP

me it really has to hit that dual channel plus minus 30 volt 10 amp is that's the main thing.

Right? Right. Yeah. And efficiency obviously doesn't matter.

No doesn't know. Actually, it doesn't matter. You're barred by physics on that one.

Yeah, it's just it's gonna get hot.

Yep. Cool. My first thing is I'm looking through the GitHub repo. And the credits are Parker, Dolman. And then Stephen hacker man, Craig.

That's right. Yep, so we'll post up the GitHub, maybe we'll get a couple of people looking at it. I don't know if anyone wants to consider stuff with it. Yeah, put it up on Slack.

Yeah, I think we need a better. I think our front panel design was a little ambitious. So I think we really simplified the front panel design. And I think if we made the board that controlled the Energon cube, make that also the digital side two. Oh, we

could probably put both of them on one board.

Yeah, put it on one board and then have a front panel that's just like the buttons. Right? And just do it that way. That'd be a lot easier.

I mean, if you wanted to do it really simple. You could have a screen that just displays what the voltage and current are. And then

just know it needs to know

the spirit member has had has still had that. So has to have the seven segment displays on mine is will do Nixie tubes then inhabit veal.

No, it was seven segment is how we had it. So that's the spirit.

But But But yeah, it did. The thing is we never actually got around to actually physically writing the code for it. Like it had like a whole country had like basic, like ability to set voltages, but we never wrote like key code or anything like that. No, no,

no, no.

One day. It's fun. So, once again, thanks, Zane for going back and listening again.

Yeah, thank you. And before we sign off, if anyone is interested in donating to my charity stream, links in the description of the podcast, or hit me up on Slack or Twitter. So thanks, everyone, come check it out on Saturday. Or help us out with SSPs should be fun.

So that was the macro fab engineering podcast. We were your hosts Steven Gregg and Parker Dolman. Take it easy. Later, everyone