This week, Riley Hall of Fictiv joins the podcast to discuss how Fictiv connects engineers and designers to job and machining shops.
The US Mint Denver produces 30 million coins a day. Denes, the tooling department manager, discusses with us how production at this scale functions.
Stephen is on the hunt for the next step in his electrical engineering career and shares the shifts in the industry and what employers are looking for.
Metrology and You
Tom Anderson
Parker is an Electrical Engineer with backgrounds in Embedded System Design and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. The following year he designed and produced an Atari 2600 video mod to allow the Atari to display a crisp, RF fuzz free picture on newer TVs. Over a thousand Atari video mods where produced by Parker from 2006 to 2011 and the mod is still made by other enthusiasts in the Atari community.
In 2006, Parker enrolled at The University of Texas at Austin as a Petroleum Engineer. After realizing electronics was his passion he switched majors in 2007 to Electrical and Computer Engineering. Following his previous background in making the Atari 2600 video mod, Parker decided to take more board layout classes and circuit design classes. Other areas of study include robotics, microcontroller theory and design, FPGA development with VHDL and Verilog, and image and signal processing with DSPs. In 2010, Parker won a Ti sponsored Launchpad programming and design contest that was held by the IEEE CS chapter at the University. Parker graduated with a BS in Electrical and Computer Engineering in the Spring of 2012.
In the Summer of 2012, Parker was hired on as an Electrical Engineer at Dynamic Perception to design and prototype new electronic products. Here, Parker learned about full product development cycles and honed his board layout skills. Seeing the difficulties in managing operations and FCC/CE compliance testing, Parker thought there had to be a better way for small electronic companies to get their product out in customer's hands.
Parker also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components.
Stephen Kraig began his electronics career by building musical oriented circuits in 2003. Stephen is an avid guitar player and, in his down time, manufactures audio electronics including guitar amplifiers, pedals, and pro audio gear. Stephen graduated with a BS in Electrical Engineering from Texas A&M University.
Special thanks to whixr over at Tymkrs for the intro and outro!
Welcome to the macro fab engineering podcast. I am your guest, Tom Anderson.
And we are hosts Parker Dolman. And Steven Craig. This is episode 293.
Tom Anderson is an old timer engineer working from HP Agilent Keysight, currently still employed employed at Keysight and volunteering at alambic, where he designs electronics for bass guitars. His career includes design of firmware, digital, analog, microwave, optical power supplies, Software and Information Systems.
Thank you so much, Tom, for coming on to our podcast.
It's great. And I have to say I'm not speaking for Keysight or Olympic today, I'm just representing myself.
Well, sounds good. So a few weeks ago, on the podcast, we were talking about taking measurements, I think we were talking about one of the devices I've designed, and, and talking about ADCs on there. And we reached out to the to the community to see if we could get a metrologist to come on. And Tom volunteered to come talk about taking measurements and accuracy and resolution and just how do you read things in your day to day work?
Yeah, most of the time, you don't really need to worry about all of this stuff. I mean, as as a hobbyist is accuracy really important. You know, well, it's up to you, because you're your hobby. But sometimes you really want things to be more accurate than they are. And so what I'm going to be talking about is, how do you make things more accurate when you need them to be. And it's not to say that you need to do all of these things. But it is good to understand them all. And I've had to over the years because of working on test equipment, you do need to worry about accuracy a lot, because that's kind of what you're selling.
Yeah, a lot of people are banking on what you're designing to tell them if the what they're designing is accurate or not.
Right. And we're banking, that they'll they'll pay a premium for accuracy, right? Because but but there really is money involved, because if you have more accurate equipment, you can actually reduce your measurement uncertainty. And by doing that you actually increase yield, because there needs to be some Guard Band between between what you measure and what your specification is, it's based on whatever your measurement uncertainty is. And so that's actually worth something. And it's definitely for sale. And then, of course, once you do measure it and you decide how accurate you need it to be, then you'd like to have a way to verify that, you know your system is better. Or that your system is at least as good as what you said it was. And so that's a whole process of verification, which is a whole nother a whole nother topic that goes with accuracy. The way I always think about it is the old saying you What is it you measure with a micrometer, you mark with chalk, and you cut with x. Right. So you're always, things always get less accurate as you go. But then if you have some good measurement equipment, after you cut with the axe, then you measure it with the micrometer, you measure what you cut. And then you carve to trim it in. So maybe that's an adjustment, or maybe that's like a pot, or maybe it's software, that's kind of the trend is to do everything in software. And so you would, you know, you carve it in and then you can start to see, okay, how close to my chalk line did I get? And so, and they're, you know, you say well is like my design uses these tolerance parts. So I could be anywhere within this chalk line. And so when I get inside there, I have some clue that I did it right. And a surprising amount of the time, you actually don't get inside that chalk mark you there's some weird error or some design flaw or whatever, where you know, I use 1% resistors. But it didn't my voltage divider didn't come out to 1% or 2% even it can it came up to 3% or some other funny number. And so that's where you have to start chasing things down and doing even more measurements. And then getting more skills as an engineer to learn about more effects. You know, maybe it's the input current of the next stage or whatever it is.
You know, that's really interesting. I haven't heard that saying before, but I really like it. And I think, I think in a lot of ways at school, we're taught just the micrometer part, we're not taught that a piece of chalk is a quarter of an inch wide. And then an axe is just going to brutalize your, your cut, right? Like, those are the things that we don't know about and what you just said there, the skill that you gain as an engineer isn't knowing the micrometer, because that's just a design parameter that either is given to you or you create, and the rest is experience in a way.
Yeah, and having equipment that's good enough, you know, like having owning a micrometer is is part of it. And if all you have is a ruler, then you don't get to, you know, make things quite as accurate, which might be fine, you know, it's up to you as to how accurate something needs to be one time I got a chance to set a specification for a circuit I was working on, I thought, I'm hot stuff, I can, I can hit a 10th of a DB with this thing, plus or minus a 10th of a dB, that's going to be my specification, because the old ones, all they all are similar products had like plus or minus one dB, and I thought, I'm gonna really hit this one. So I built it, and they put it into the production test chambers and so forth, the environmental test and all that. And I came in one morning, and the production guy said, Hey, your circuits, it's, it's failing the, the gain test. I said, Well, how does that and he said, well, the specification is point one dB. And the measurement error is point o six dB. So your part is point o four dB. And we're measuring point o five, what are you going to do about it? I thought oh, man, oh, I wasn't thinking about measurement uncertainty when I made that brag, right. And so actually did fix it. It was a pretty, it was a really subtle problem, like you're saying it was I had to learn about a whole different effect. And it was a shielding problem where one stage was feeding back to the previous stage a little bit and just changing the game at a different setting. So the game, the gain of the previous stage depended on the gain of the next stage, when the gain of the next stage went up, it felt a little bit of that signal back to the previous stage, and made them so that they wouldn't add together anymore more accurately. And so. So that was that was a problem. And so this fix was to add more shielding, which we did. I don't know actually, what spec it finally shipped with. I'd moved on by that point. But But yeah, that was a really tough spec. That was in a spectrum analyzer. And I which is something I spent a lot a long time working on.
Yeah, it's it's easy to set the spec without thinking of some of the downstream consequences of doing well. That's
that's the easy part. Right? I can write a spec for my next product in like 30 seconds, right.
Also, I like that. I'd like that mentality you had there? Because I mean, I'm sure a lot of us have been there before. It was like, I'm going to be the one engineer. I'm going to be the guy that gets this. It's like no, you're not. No, you're not.
Yeah, what are the phases of an idea? It's something like, wow, this is really simple. I don't know why everyone doesn't do it this way. You know, what's this is? And then you try it. And you say, Wow, this is really pretty innovative. But there's these problems. And then it's like, oh, this is why nobody does it that way.
Yeah. Yeah. I'm back at Bluebell while saying, usually actually, it's not the engineer setting the specs, either. It's usually like a sales
or marketing team.
Yeah, and in this particular product, the project manager when it's kind of traditional, when you go to start working on a project, the project manager tells you what the requirements would be and, you know, you kind of sign up for it to see if you're up for such a thing. And so, so, this was a very unusual project that was cost driven. Unusual for it was HP at the time, and and he said, Well, this is what your assembly needs to cost and he showed me a spreadsheet and and how that contributed to the overall product and all have that and I said, Well, what does it need to do? And he said, well, it needs to cost this amount. And I said, No, but like the electrical specifications. He's like, Yeah, whatever we had before, just just make it like that, but but make it cheaper. And so my first step was to actually measure the old product to see what it really did. And by the time I was done, that they did a whole bunch of production change orders to fix what's wrong with the old product that I uncovered in like, actually analyzing exactly how well it worked, which, which they had done some of but not quite enough for my taste, or certainly not enough to, you know, base a new design on it as a specification. So
you had to maximize everything else, but minimize the cost.
Yes, it's actually a really interesting problem. If you analyze it that way, the most important thing is to be able to model your cost. So to be able to, as you're designing to see the impact of your decisions on the final cost. And that's actually a hard problem. Some people I think, have tried to make CAD plugins and stuff to work like with contract manufacturers to predict what something's going to cost kind of as you go. So you have a little cost in the corner of your, of your CAD tool, like a plugin and a CAD tool. And that's a great idea. I didn't have that. But what I found, and that doesn't really guide you for electrical decisions, like what component do I use, but what I found worked really well was actually to minimize power. And that really indirectly minimizes cost in a very strange way. Which is that if you, if you minimize power, you can make things smaller, and they won't overheat. And as you make things smaller, you use less material, you get more boards per panel. And a lot of the cost driver is actually your boards per panel. And so if you can reduce that, and get everything on to smaller boards, the cost goes way down in these big steps. As you know, as you get one more or two more, three more on to the panel, you can see this sort of staircase, of the cost going down, and it's one of the few things that, you know, you can really do something about. And so really sighs You know, in this case, putting components on both sides of the board actually made it cheaper, because it was, you know, fewer panels because there was this big panel, at the time, the manufacturing process we were using had a large panel, per panel cost driver. And so, but it was important that I knew what those cost drivers were. And and so that's one of the challenges with a lot of manufacturing, is that like a contract manufacturer may not want to share all their cost drivers because that's some proprietary thing, right? That's kind of their secret sauce as to exactly
what everything costs. Is that secret spreadsheet?
Yeah, yeah, exactly. And so yeah, and then you have people like submit like 100 Fake designs so that they can reverse engineer your spreadsheet so they can figure out what your costs are either because they're another contract manufacturer and they want to compete or because they're trying to design something. So ultimately, I think the word gets out I think more of the challenges that cost drivers keep changing as you get different machines or your processes change or you move factories or something like that. It's kind of hard to design that way unless you really know what exact factory you're aiming at.
Funny, funny little side tangent here, that's right along with that. Years ago, I think it was before I started at macro fab, I found a website that allows you to type in some generic board information X Y, size, color, things like that. And somebody had gone out and reverse engineered the algorithms of like 30 Different PCB manufacturers. And and they would they would give you an aggregate of everybody. And it was awesome. It was great if you were you know, hobbyist level trying to get you know, cheapest possible price with with information. But I'm trying to save that one pint of beer, that one boy, that one point of view, but it's just like oh my there and the way they even set it on this website the way they did it was just putting in loads of information into their calculators and extrapolating data and then curve fitting to do you know, I don't know how often people change that information was like, man, your website is valid for a week. You know?
Yeah, that's rough. That's well that's kind of Like the a lot of these big data problems where, you know, you need you, yeah, there's this really nice computer programming language for doing statistics called R. And it has all this really advanced curve fitting stuff in it, if you ever want to take out super deep dive into curve fitting, I really recommend getting into are because you can take in, you know, these great clouds of points and turn it into like, well, from this big cloud of points predict me this cost. And it can make you a model based on a lot of things. I did that actually with spectrum analyzers when I was making spectrum analyzers. And I've found out that there were all these different terms like how you could predict based on the datasheet for a spectrum analyzer, what it would cause one was just how many watts did it take. So that's getting back into the power thing. Another was the maximum frequency, and I found that there was a term, I think it was something like 14 cents per root hertz. So you take the square root of the frequency and multiply by 14 cents, and that's like an add on to the cost. And then there was a few other terms, you know, like the length and width and height of it, or the volume, actually, or density is what it actually was, because, because ultimately, if it was, if it's made of like a solid chunk of aluminum, it's more expensive, because what that means is there's a bunch of microcircuits in it. And so, because that's how they used to make a product that the goldbrick microcircuits use product. Yeah, yeah. If you put lead in, it's like, yeah, I had, I've seen that in telephones, the old desk telephones that, that were would be super lightweight, and they would fall off the desk. And they had weights in them. But anyway, yeah, you can you can predict the cost of things in ways that you wouldn't expect. With printed circuit boards, of course, though, I mean, all the main, there's so many manufacturing steps. That of course, it makes sense that there would be cost drivers, but they're pretty complicated, because there's a lot of steps for bear boards, and then even more for loaded. So
there's a lot of steps and there's Oh, what are we at like 60 years of just different technologies and different kinds of machines, that kind of stuff that are making them now?
Yeah, on so many different substrates and, and materials? Yeah, it's, it's a, it's a complicated thing, for sure.
And project managers must have loved you given them little calculators to just say, here's our next product. And here's how much I expect it to go.
They had, they had no, they had absolutely no use for this. The only use I ever put to this was I actually came up with this when a new spectrum analyzer came out. And we were guessing what it would cost. And so I did this curve fit. And I said, Okay, well, we got to have a pool, and whoever wins the pool, the other guy's got to chip in and get a free lunch at the local microbrewery. And so, so I put together this model, and I came up with this cost. And I won for sure. I was within like, 3%. That's bad. Yeah, it was kind of spooky, actually. Yeah. Yeah, I actually learned the technique from my dad, they had something like that, for aerospace, actually how much wood or radar cost? I think with the thing, and they could do real high level analysis it when you get into systems engineering, because system engineering is really about, you know, if you put a lot of things together, are they going to work at the end? And so if you take a system, like an aerospace system, that's a huge system, right? You It's ungodly expensive to, to put it all together, and then say, well, yeah, we're gonna light a match under this thing, and it's gonna go to the moon or whatever. Right? So, so the system engineering that goes into that is incredible. Whereas, as products get smaller and smaller, the role of the system engineer, you know, it's smaller and smaller. When you're at a single board product, it's, you're pretty much just the board engineer, maybe, you know, the person who tells the board engineer to talk to the me, it's the system engineer. So but in that, you know, in the region where you have a box of with a dozen boards in it or something like that. Yeah, you need a little bit of system engineering to say, you know, how are you going to put things together? And that, of course, gets it like we were talking about before it gets in the specifications and the accuracy. So with that great segue there. What what you're worried about, because I have done a little bit of system engineering and And I'm not I wouldn't count myself as a metrologist. I never had a business card that said that. But, but I do work with metrologists. And I know about what they do. And I know what they worry about. And one of the big things they worry about is temperature. Because just about everything is a thermometer. Right? And so what do you do about that? And how does that affect what you put together
first? On the first page in the manual, it says set air conditioners 72 degrees?
Yeah, there's definitely like, yeah, that's, that's a really good point. Now, so when you have an air conditioner, it's got some high temperature on it, and a low temperature, right, because it's got some window, where it's, it's ramping down the temperature and then as it's on, and then it turns off, and the temperature creeps back up. So you get this kind of sawtooth pattern. And then as people open and close doors and stuff, you get these little steps, and you get little gusts of wind. And then if you like wave your hand in front of a circuit, it'll make a little breeze, that'll cool things down or warm them depending on how the air is flowing. And you can see all of this in a in a sensitive enough circuit, you know, you can just breathe on him and see it. If you if you turn the measurement way up. Now part of it is getting that measurement to be sensitive enough. And so what you got to do to do that is to figure out how to really kind of zoom in with your test equipment, because I see when a lot of people use test equipment, say they're measuring something like rise time, I'll just see a step on the on the scope. And they said, Well, I hit the rise time button. And it tells me that it's you know, two nanoseconds. And so I said, Well, I see a step there. It's like the points are right next to each other on the screen, how what's the resolution is it like one nanosecond resolution. And so what you need to do is to zoom in on that rise. So that and then, so that it's spread out in time, maybe the rise time, you want to go all the way across the scope, like from the first division to the 10th division, you want to see that, that rise, going up slowly, and you want it to fill the whole screen and the scope. And so it's it kind of banked to a 45 degree angle. And then you put the markers on it, and you really measure like 10% to 90% or 20 to 80 depending on what your specification is. So you really get that that measurement, right. And now if you see that, and when you zoom in, but you see that it's like bouncing all around. You say well, wait a minute now, is this rise time with respect to this clock? I mean, what do I really care about? Is it the threshold that the logic high threshold after the clock pulse? Or how am I synced up and this measurement, right? It's, there's, so if you know if it's rock solid, that's great. You don't have to worry about all that. But you do get into this, this whole jitter problem. And jitter makes it very difficult to, you know, understand. There's very little mathematical analysis, that none of that they teach you in school, where you worry about time shifting a little bit, because they always say, Well, sine wave, it's sine, you know, two pi f t, right? They don't say, Well, what if T has like a little noise on it, then I'll talk about that very much. They'll talk about it the other direction. They'll say, Well, what if the amplitude has a little noise on it? And they'll spend a whole semester on that. But time noise is a no that doesn't exist, that's too hard. We don't know what that is. But it's it's scary, but it can be analyzed. It's expensive to work on. But anyway, when you so when you're measuring making this measurement and say you get a really good high resolution measurement, and you have temperature, but it's cycling up and down what that's not great because you don't really know where you are in all of that cycling of a couple of degrees it's very, very difficult to make a room that's plus or minus one degrees C worst case to do that, what they do is they actually make a special temperature control room with many temperature sensors. And then they have coolant in the walls that circulates through it and then the air is pre chilled and then heated with a proportional heater, not a not a on off like an air conditioner but you know just a proportional control heater. And then you you duct that in and that's what's that's how you make a metrology room. And you know you have to worry about humidity and all that stuff to a little bit. And kind of, they kind of tend to be kept kind of clean. They're not quite clean rooms, but they're pretty clean. Like you won't see people eating in there. No burritos, you know.
So on your on your test equipment.
Yeah. Yeah. So anyway, that's sort of so that's electrical, and mechanical metrology, no mechanical metrology, they get into like the coordinate measurement machines and that sort of thing. And the, the way that I got into it was because in microwave standards, the, the standards are based on mechanical measurement. It's like a length of a transmission line or something is the is the standard. And so you translate this mechanical measurement into an electrical measurement through a model. And you see a lot of that in accuracy, things where it requires a model of the system in order to say how accurate something is. Another way to say it is, if you want to say how big something is, you have to say what shape it is. So if you want to ask me that makes sense. The dimension of this is well, is it a rectangle or a trapezoid? Or what is it that you were trying to make here? So that's the model. And so the accuracy works on the model. Yeah, so with temperature, say, there's a lot of, you can make really great measurements with frequency, right, because we have incredible resolution on frequency with crystal oscillators and stuff, you know, and all the work that people have done with really fancy, like even a cheap function generator you buy these days has beautiful synthesized frequency, it's really accurate. And if you have 10 megahertz references in the back of a lot of these pieces of equipment, you hook them together, and they're all in tune. And so, if you make yourself an oscillator, either a crystal or an LC oscillator, and you measure it with a frequency counter, or spectrum analyzer, whatever you have, you can, you can really see it drift with incredible resolution. And that's where you can make yourself you know, milli hertz resolution thermometer, you have no, it's not calibrated, you have no idea what it does. Well, I suppose you could calibrate it if you had a good thermometer. But that's where you can see these little waves of air go by and that sort of thing. And you can tell somebody's got that problem when you go by their bench and they have a paper coffee cup on top of their circuit. That's when they realized that the drafts in the room are pulling the frequency around to their measurement. And it's really, it's amazing how well it works. Just for first approximation, frequency stability coffee cup works every time.
I remember doing something similar at a macro fab, I had to have some, I was matching some transistors. And I'm for a frequency based circuit. And I was just for fun doing hand match transistors, and I got a big block of Styrofoam, put all the test transistors in that and put a fan on it to make sure that it had constant, cool air flowing over them to try to make as much of a specific temperature as possible, or just to make sure that random fluctuations were not included in that it still was incredibly difficult to do.
Yeah, and you get into microphonics as well, when you start adding fans, and another feature of well, DC fans anyway, is the rotating magnets. They tend to cause all kinds of interference. Plus they have a very strong pulsing of their current like you put them on 12 volts or whatever your 12 volt supply, it gets these big pulses of current through it. And you know, that's why you'll see systems sometimes have their a separate power supply just for the fan is because they don't want those pulses of current coming in and polluting their supply. That's another another big source of error is do you really know Mr. System Engineer, where all the ground current is flowing in your system? Because it's a really hard problem to know. Like if you have like, say a box of cards of electronics with a metal frame and there's kind of some grounds sprinkled around. Do you really know how the chrome current is flowing? And most of the time unless you think really hard about it, you don't? Another thing that's really good to have as my groan meter, or milliohm meter, a meter that can show like 1 million full scale. Because you really need to know how much resistance is there from this frame part of the frame to that part of the frame or across this ground plane, or whatever. Because here's what happens. Let's say you've got I'll make an audio connector. Yeah, let's say there's an audio connector on the front panel. And it's, it's on the casting, you know, and it's got, and it's hooked to a little circuit board, this a little amplifier. And the little amplifier has a power supply on it. And the power supply has got a nice red wire and the black wire and the black wires, the ground and the red wires the plus going into the supply. And what you're expecting is that current flows through the red wire and then through your circuit and then back through the black wire. But there's a very good chance that instead of that black wire is just this beefy little wire, and it's got this really low resistance frame that you attach to the quarter inch audio input ground to. And so that return current is going to go through the frame, and then back to where your supply is mounted on the power entry to you know, to it same as your safety ground of, of the whole metal box. And so you now you have this big loop, right, that's, that's the red wire with a plus 12 through the circuit, and then now back through the frame. And so any magnetic field that goes through that loop is now going to be imposed onto that into that. And so it it's going to lift a voltage somewhere. And there's a very good chance you'll be able to hear oh,
I've definitely come across that. And it was 60 Hertz. Really hard in that circuit. Oh, yeah.
Yeah. And actually that, that I figured it out
by accidentally disconnecting the ground, the ground wire. Mm hmm. And that a 60 hertz went away. And I'm like, Ha, because it was flowing through the chassis, right, took off the loop, right accident.
And sometimes you need to make the loop better. It's like, well, the loop is good. But maybe that chassis isn't really designed to conduct reliably all the time, maybe it has some paint on it or some other finish that's not, you know, chromate, whatever. And so, a really good tool is an exacto knife. So you go around shorting things together with an exacto knife, it makes a really low contact resistance because you want to dig into the metal and get to that good fresh metal that's not oxidized. So you want to dig through that oxide layer and cut in. And when you do that, you get to you know, you get to make these low contact resistance and you can just listen or watch the signal on an analyzer or whatever, and you can see it go down. Another one is like torquing screws, it's like I've been able to see like a signal on his screen just drop as I added more torque to the screw until it just gets to the screw yield point and it's like oh, I should I should probably stop now. I'm about to pop this screw
Yeah, nothing tighter than stripped I'm envisioning a an exacto blade with a banana connector on the end of it. Just go to Rod stab and
it's a blade to blade connection that's really low resistance. So you want to you want to cut in both ends now if you had to exact those with a cable between them that would be pretty good too but it'd be need to be like a jumper cable on the car or something like that. And you need to get the blade attached to that cable really well. So you can imagine like a big copper tube but you know you really want copper wire because even like Plumbing copper isn't really very good. Do you ever measure like the resistance of like copper pipe it's not that good. It's not nearly as good as the like electrical copper anyway.
So I think most of that is from the because I have messed with that before and it's I think it's actually the the patina the oxide that forms on top of copper pipes is what makes that super high resistance.
I think there's some oxygen involved as well. Oxygen is
there more oxygen in the copper and pipes? I think
so but I don't have any numbers on it. I've never had to actually use plumbing supplies in and in a product for electrical but I I could just imagine like those, you know, you see these big data center, bus bars and stuff you don't where they have these huge currents because the trend is now of course, towards all these super high current low voltage things, you know, like an FPGA with a 10th of a volt supply 80 amps or something. You know, there's these super high current low voltage. It's it. It's basically we're making arc welders
as nanometer scale.
Yeah, that's really, really bizarre. So, so but anyway,
on that. Sorry, Tom. So on that real quick, is that from the reduce on on rise? Is that I think that's the reason why going lower voltage is like the future. Future. Why,
why is lower voltage better? You mean? Well, there's a, there's a couple of really good trends for lower voltage. One is that if you have the same volts per nanosecond, you get there quicker. Just because you don't have as many volts to go. So if you look at the rise time of modern stuff, it's not that different than it was a long time ago. It's just that we had to go five volts. And, you know, so it's five times
faster. That's why I figured I just the other is the one energy. Yeah,
there's the one half CV squared term, and that has the V squared term in it. And so that's, that's a really big deal, too. As the device count goes way up, the energy on each device, that one half CV squared, Dart use, pretty soon you're talking about real energy with a billion transistors, you know. So that's, that's a big part of it, too.
Sorry to interrupt, then. No. I was curious. I, I thought that's why I remembered. So
yeah, with the smaller breakdown voltage you can get, you know, I think they can do lower capacitance, just smaller devices, yeah, smaller devices, you end up with a smaller breakdown voltage. So you have to run lower voltage also, that's the other thing is you don't really have a choice. For the small devices. It's the volts per meter or whatever. Yeah, if you're looking at the volts per meter and things, it's kind of scary, actually, like if you just look at two traces on a circuit board, and you say, like, well, I've got my plus and minus 15 volt supplies here. And they're right next to each other on these traces. And they're, they're separated by five mils. That's five mils is 1/200 of an inch. And you're saying you have 30 volts across there, that's 30 times 200, that's 6000 volts per inch. And that's getting dangerously close? Well, that's getting within the region where you should start thinking about what you're doing. Because even if you don't get arcing, you're going to promote corrosion there. Right? Any any salt or any moisture on the board, and you're going to start growing dendrites and all that. So yeah, another good thing about to look at for board design in general, is you definitely want to think about your volts per meter. As far as I know, there aren't great cad tools for that. Do you guys know of any tools for kind of checking a layout for volts per meter?
No, no, besides just manually calculating out when you're doing your layout, right? I imagined your I met I imagined cadence probably has a package for that. Because they have a package for everything. Oh, there you plugin. Yeah,
that's an opportunity for someone.
So I want to take a quick step backwards on this kind of a little bit back to taking measurements and actually the act of taking measurements. And one one thing that's sort of been a theme of what we're talking about here is when you're going to take a measurement, it isn't as simple as just going and grabbing the measurement, you have to first start by asking yourself, What am I actually measuring? And what is my goal with this measurement? It's not necessarily that you're, you know, what the measurements should be? It's that you need to know how you're attacking it. So like we said, what, what temperature Are you are you maintaining temperature properly?
I mean, all the bazillion of other factors that goes into it. But what I want to talk about here for a second is let's just pick something simple that electrical guys do. Let's just talk about voltage measurements.
And say I've got Joe Schmo voltage meter. And I'm needing to tell say a test production and engineer measure voltage at this particular place, it's really easy to say, put your red wire here, or maybe I even thought about it and gave you a test point on the board. But when it comes down to measuring against the ground, what what is the kind of target with that? How do we I guess what I'm getting at is? Where do you put your ground? Good example, you got a circuit that has some analog stuff in one corner, some digital stuff in another corner, some power supply entry in on some other side of the board? Are you trying to measure your ground closest to the analog stuff? Are you getting the correct reading there? Are you putting your ground over at the digital side of the board? And you reading something else? Which one's right? How do you go about approaching that?
Yeah, well, the short answer is you only get one ground on your board. And you have to decide where it is. And everything else is you can call it ground. But you're hurting yourself when you do it. Because it's not really grounded, grounded, grounded, spot and offset. And it's like, well, you can give me your different grounds a little hats for them to wear, you know, you can make them Oh, this is my digital ground. You know. Here's my analog ground. And here's my earth ground. And you know, I think there's somebody wrote some amusing cartoons with a whole bunch of different grounds on it. So what are you going to do with all of that? Well, usually you want a stable reading. And so if the ground that you're measuring has is, you know, like, there's probably a capacitor between the voltage that you're measuring, and the and the ground, like, maybe you're measuring the supply of an op amp. And so in that case, you would want to measure right across the plus minus terminals of the op amp. Or if you're measuring the output voltage of the op amp, you probably want to measure someplace close to the op amp, as opposed to off in the corner, by your switching power supply, or your digital section. Because you want a nice stable reading, right? That's the whole goal, you got to ask yourself, why do I care what that voltage is. And so if what you really want to know is, well, how bad is my ground, then what you do is you measure from ground to ground. And, and you may be horrified at that, well, you'll at least find out what's going on. The other nice thing about measuring from ground to ground as you can turn the sensitivity of the meter way up. So you can expand it like I was talking before about the scope about how you should always, you know, expand the scale. Well, if I crank the scope all the way up. And I'm I measure my ground point. Either either that ground point this far away or close by, I should see zero there. If I don't see zero there, then maybe it's a little loop, I have a little antenna, that's the loop of the probes. And maybe I want probes that are more better shielded. And maybe a smaller loop area, like a like a little piece of coax with just a little wire sticking out the end. And maybe I want my ground test point to be right next to my voltage that I want to measure. And maybe have, you know the little spring clip that comes with your scope probes, you know, that little plastic bag that comes with the probes and it's got that little spring in it along with the little clips for the different colors you ever use that spring or do to set that, you just throw that thing aside, that little spring, you take off the bayonet tip, and you jam that spring on there. And what it's for is so that you can get a really small loop in the scope probe between that the tip and the ground. And you can get much it cleans up a lot of problems with scopes when you when you use that little clip, but you need to have a ground that's nearby. And so one thing I always look for in a layout is Do I have a ground near the sensitive node that I want to probe because that's the spot that I want, whether it's with a voltmeter or scope or anything else. And if I'm fancy, what I might do is actually put a connector there that's normally a do not place because that way I can get a shielded connection on to it maybe and I can get a really accurate measurement then. If I have room, I'll put that in and they have these great little surface mount. The things that go inside of like Wi Fi modules and stuff that have the that little tiny, horrible surface mount connector that everyone hates.
Yeah, it's smaller than an SMA.
Oh waistline Oh yeah, cuz it doesn't have any snap on ones. Yeah. Yeah, they're horrible like a you they rip off the board really easily? Oh yeah, they're terrible. You wouldn't want them in production, but they're perfect. Well, unless you're only going to click it once and like go into your product, IoT antenna or something like that they're good for that. But they're good vis test points.
Yeah, actually, I've just, that'd be awesome, actually. Because they're not that expensive, either. That's right, yeah. And they're surface mount. So they're easy to place. Oh, that's a good idea.
Yeah, and you can get a cable that goes to him, and then just get a cable that has that on both ends where it goes to SMA and then you just cut that cable and mount it to whatever you want. Put it in your scope probe or whatever. The problem is there only that cable is probably only good for like 100 cycles. So it's really more about diagnosing something for verifying the design rather than for a production test. You could use it for production tests, you just have to have a lot of those cables lying around. They make actually they make us there's one company I think that makes a stainless steel version of that. It's an adapter that you can get the stainless steel and it actually does last a long time. So it'll the connectors will wear out. But you don't care about that you're only going to use it once or twice.
And BX I think that's it.
That's one of them. Yeah, there's a bunch of little tiny ones like that surface mount thing. They're awful. Yeah. Yeah, I have, I've actually
run into issues with what we were talking about there. Just a moment ago, where I've had test a test engineers call me up and say, Hey, we've got this new batch of devices. And I'm turning this calibration trim point here. And we're just we're not in spec with everything, everything's outside of spec. And I go down and check out what they're doing. And they've got, they're, they're measuring their test points correctly. But they've got their ground plug all the way over at the side of the plugged into the power supply. And it's got cabling to the device. And so their ground is is forever huge away. And, you know, you bring it over and, you know, way more in spec.
Yeah, helps a lot. And just having solid connections helps a lot. I mean, you know, the classic thing is, I think, I think this happens to companies, like when there's a new guy, you take your, you open up your drawer and get your broken scope probes out. And then you go over to the new guys area, and you swap your probes with the brand new probes from his and say, yeah. So I'm not sure if we still do hazing of new engineers, but it's probably frowned upon. But yeah, you really need to have a way to check that your probe is working. And measuring ground is definitely one of them. It's also nice to you know, hit a supply. And I always say check the supplies first. Right? Because you can chase your tail forever
with that noise might be coming off your supply. Well, not only
the noise, but this is the supply even on is it oscillating? Is it whatever? Yeah, yeah, my dad was a radio amateur. And he bought a receiver one time, and he's it just it, he could get signals on it. But it was just really bad. It wasn't sensitive at all. And so he starts tweaking the front end and you know, working on the shielding, and he's getting, you know, making filters and stuff to try to clean it up. And then he discovers that one of it was a tube thing one of the tubes the power supply wasn't even connected. It had never even been soldered. It just the radio had never worked. So so always check the supply first. Because the signal made it through the tube, it just didn't amplify it was an attenuator instead. Yeah, always always,
what's the saying? Thou shalt check voltages? Right?
I think that's good. Good. Yeah, yeah, it's and it's easy to check voltages. It's, it's easy to do.
Yeah, you were talking about like, what you care about. When you when you go into for testing, it's similar in God do a lot of automotive stuff. So like, most time I'm just talking doesn't have power because most time on cars, it's 12 volts or five volts sometimes but most times, so you have just have a test flight. Right? And so you're just poking stuff and having a light bulb light up. Then you go okay, that's working. Then you go okay, I need my multimeter and figure out am I actually good Adding my full 12 volts at this point. And then if you go in further because they're trying to diagnose electrical problems, then you get this oscilloscope and plug that out in your car seat and start seeing Okay, is my signal clean and Oh, my voltage clean enough now.
Yeah, and the test light actually has the other benefit that it not only checks the the voltage, but you can see that that voltage can deliver a little bit of current at least. Because one of the things that can fool you about a meter is you know, they pride themselves on having really high input impedance, right, very low current needs to flow. And so you can have something like a dead nine volt battery, and it measures fine. But if you put like 100k or 10k resistor across it, you can see the voltage sags down right away. And that's because it can't, yeah, there's a voltage, but it can't deliver any current. So I like the test light a lot. And you can kind of look at it and see is that a good 12? volts?
Yeah, you can see how bright the light is.
Yeah. It's actually quite nice
time when you're dealing with electrical problems. It's bad grounds. And you won't see that typically with a meter, you'll see that with your tests, like because you're trying to pull some current through that bad ground.
Right, right, right. And then you turn the thing on or whatever. And then you see the light dim, you know that?
There's, that's like, on the other end of the spectrum, we're talking about accuracy. And that's like, using your eyeballs to figure out how bright the light bulb is?
Well, it gets into how, yeah, and and what are you really trying to do? And and that's actually an even harder problem. I think you're trying to make a system work. And so what does work mean? And that, that becomes, it's always nice when you have specifications that hopefully you got from somewhere. Maybe you have a marketing department, and they helped you.
There's some danger in that too, because the marketing department may have just downloaded the data sheet and said, look, the front page says that this IC can do this. So here you go.
Yeah, typically, it does something like that. Yeah, the problem with data sheets is like sort of reflect the aspirations of the marketing department of the IC company.
Aspirations Yeah, that's a good word.
And, and really, also,
when talking about accuracy, I think the car guys, when they start actually getting into numbers, that's where the accuracy starts raising eyebrows because it's like, Hey, I'm trying to read this 12 volt rail, it's 10 to 14 volts. Good enough. It's there, right?
Yeah, you get, you get to know your battery pretty well. And 10 volt batteries and cars are not in good shape. But they have an incredible range. I mean, automotive supplies are famous for for, there was a great linear tech datasheet sort of boilerplate that was in their data sheets about automotive power supplies that started off with you know, this is the power supply from hell. And it went through, you know, all the different scenarios including jumpstarting double jumpstarting where they jumpstart a car with 24 volts instead of 12 the spikes that go through automotive you know, 65 volts spikes everywhere. It's, it's a very harsh environment. And so it doesn't surprise me that well, actually, I'm amazed that it works as well as it does. They must spend a lot of time working on you know, like, power supplies and voltage spikes and stuff because I've never worked on cars, designs, that way, it'd be really hard.
It's gotten a lot better than in the past. Because nowadays, your your high voltage stuff, which is your spark plugs, your ignition coils, usually, nowadays piggyback right on the spark plug. And so the distance for the high voltage is like the connector socketing onto the spark plug. We're back in the day your you had one coil and usually was like mounted like on your firewall, and you had like five feet of cabling going into your distributor and then that spread all over the spark plugs. So you had this huge, you know, 30,000 volt events spread apart all over your engine and then one of those sorts, like arcing to your ground and yeah, it's got a lot better nowadays. But yeah, it's it's interesting that it works. Especially when you talk about like, it's 12 volts but it's actually really like 13 and a half to 14.
Right.
And you're in the US like the battery as like the capacitor in the circuit. Yeah,
batteries, it helps smooth out is a pretty good capacitor, you know, but but people would complain, you know, if you used it as a capacitor, people would complain that it wasn't very linear. You know, because, but you didn't model it. If you model it as a capacitor, when you when you're first discharging it, it acts just like a capacitor, it has so many cool ohms in it. And it's, like a lot of things. linear systems are great, right? Because if you look at a small variation of things, almost everything is linear. And so I talk a lot about linear systems and, and batteries make it great when you can calculate how many ferrets in your battery by just looking at, you know, how many amps can it provide, you know, for how long and so that was that many cool ohms and that much delta V and q equals CV and it all comes out. And it's an excellent exercise, it comes out to a really big capacitor with a really big manufacturer ads. And, and very low resistance. So it's very high performance. The the ones that are driving me crazy right now are actually multi layer chip capacitors. They have this horrible problem with the voltage coefficient, have you guys run into this yet? Where you increase the voltage on the on the capacitor and the capacitance goes down, and not by a small amount, you can easily be down to you know, 20 30% of your original value. And so that's a that's a tough one to figure out there. Fortunately, some of the manufacturers have curves that show you where you are, come in and Mirada both have curves. For most Yeah,
they're, it's surprisingly actually those curves are not in most data sheets like it's really I've seen them only on Kemet and mirages, most other manufacturers tend to just ignore that or not have that on their data sheet.
Yeah, yeah. And then for aging, that's another problem with these particularly high density parts. as they age, the capacitance goes down, to the extent where I think you would probably see it like deer, when, when you're loading parts, some of the like pick and place machines have a little like capacitance meter or something in there to tell if you're loading the right value. And but some of these parts of age so quickly that you'd actually have to put in extra window extra margin to keep them from being out of spec, just because their capacitance is going down, you know, so much per month. So you'd have to know the age of the reel, in order to, to know, you know, me just sitting on the shelf, yeah, just sitting on the shelf, the capacitance goes down over time. And certainly with during reflow, too, well, during reflow, the capacitance actually goes back up, because it reaches like the Curie temperature of the the barium titanate or whatever the heck it is in there, and then it and then it. And then as it ages, the capacitance goes back down quite quickly, in the first day, supposedly, it's pretty unstable. And then it settles down and, and it goes goes down for quite a while. It's proportional to the log of time supposedly. So you know, the between one between one day and 10 days is the same as from 10 days to 100 days, 100 days to 1000 days and so on. You got a certain percent loss. But apparently, I just got information recently that for a capacitor that's held at its bias voltage, it actually ages even faster. Which we're not seeing in the data sheets. Except, except for a very small disclaimer somewhere that says yeah, it goes faster if you put voltage on it. So yet more capacitance, Lawson none of these seem to increase the capacitance, except for reflow. That helps.
Yeah. So in a decade, we've gone to all sorted reflowing all our ceramic capacitors again, you know, all our stuffs to start working again.
What it says as the electrolyte dries up, they can increase in capacitance.
And well, that's thing is refurbishing old electronics. just replacing electrolytic has been a thing like since like radios in the 50s and stuff like that. But ceramic capacitors is something that no I haven't thought of having to replace in the future because it's like, it doesn't have anything to dry out. But it does age. As Tom says,
yeah, it may be that, you know, most of the bypass capacitors we use are way larger than we need. Need, it may be fine. Maybe we should keep doing that.
Yeah, may point one, maybe there's a reason for the ginormous value of point one microfarad.
Yeah, that's a pretty, it's a pretty big value. But, you know, they make these really high density, small parts. Now you can get like a 10 Micro ferrets and, and Oh 603 or something like that. And these really high density parts, you know, fairly decent voltage, I think 25 volt. Oh, 603 10 micro farad. It's, I think that's a thing. I don't know if you can get them anymore. But I'm not sure how good they were. And I'm not sure that by the time you'd put down these aging factors and all that stuff, you might have been better off with a one micro farad. And knowing actually having a slightly better idea, or point one micro farad even better, knowing what its value would be in the future. So yeah, use use a more conservative value might be okay, or a smaller value. In general, when you push the when you push the capacitors to say I want the most capacitance, I can get in this package at some voltage, and you pick that one that breakpoint in the datasheet the edge usually a mistake. Yeah, don't do that. You gotta go down a couple of values at least. And in the case of these high density ceramics, you may need to go down way down the page.
Speaking of like this aging stuff, so there's a group of people who this is this is getting audio stuff like stompboxes and effects pedals, where people care about the how old the batteries are, that drive these circuits. Because apparently that changes what how they sound. I can totally see in the future marketing, that this stop pedal is like nine years old, and it's ceramic capacitors are like aged just the right way now
Oh, yeah, yeah, you might be able to hear that. Yeah, it's amazing what you can hear. I've been working with the Olympic on on bass guitar, circuit senate is amazing what people can hear. You know, you hear about all kinds of gold near stuff and all kinds of things. But if you know what to listen for, there's a lot of the problems are really just bad. I mean, they're, you can you can really hear it, even if you can't hear it. If something changes it, you can tell that something has changed or is changing. An example would be if you play through a compressor, you play a guitar through a compressor, what it's doing is it's changing the game, while you're playing and adapting the game to be a more optimum level based on the way you set the knobs. And so you say, well, that's fine, that should work great. But at the same time that it's addressing adjusting the game for your signal, it's also adjusting the game for the noise in the channel. And you may not have noticed the noise before. But when the noise starts going up and down, you can hear it. And so it wasn't something you could normally hear. But the change in it is something you can hear. And so that's that's where I think a lot of the sort of weird effects of weird things that people can hear come from this sort of, well, it changes in a different way. And maybe they think they tracked it down to the battery or something, you know, maybe they're wrong, and it wasn't the battery, it was something else that they changed at the same time. It's very difficult to do these experiments because your ears get tired. It's hard to you know, if if your test equipment is your ears, they're not particular. They're kind of suggestible, if people tell you things, you can kind of hear it. And they're also, you know, they're not real accurate. Over time,
like listening to static and thinking you hear ghosts.
Yeah, yeah, sure. Yeah. Hearing hearing signals out of the noise. Yeah, yeah. And even like, what is static, right? Like, there's thermal noise, that we kind of agree. We all agree that static but then there's kind of the sound of machines that you hear from like, a radio too close to a computer, you know, that sort of buzz and wine that you hear? Well, if you know what to listen for, you can hear different things you like I know a data bus from an address bus. I can listen to it, I can tell you which one it is. And if there's if sometimes clocks, you know if there's more than like, if there's more than one clock and their frequencies are shifting a little bit that will cause some mixing product that you can hear. There's all kinds of things Things that you can hear that, you know, are kind of in that category of noise like, and they're not particularly musical, but you can learn to hear them. clicks and pops are the worst, like, most, you know, as a hobbyist sort of making a stump pedal or something like that. The worst thing when you first get started as well, yeah, it's a great effect. But when I turn it on, it makes this big popping sound. Right? So how do I deal with that? Right? Well, it's not, it's not really a steady state problem, you know how to it's, it's a, it's a hard one to do. So yeah, sort of hiding, hiding these things from the human ear is difficult, because it's, there's, it's actually pretty good at a lot of things. It's hard to explain. So I guess I'm simply goldenear sympathetic, but not, but not so much that I would pay a lot of money for it. Unless I knew what I was buying. And I believed it.
I really liked that I might have to steal that golden years sympathetic.
Well, you can really do a B testing. You know, I think you did that on a project before with capacitors, right? Could you hear these capacitors? And which one? Sounds better? Different strings?
Yeah. And you can? Yeah, it was a terrible circuit design. wasn't that bad of circuits? But it wasn't optimal? For sure. Yeah, you could definitely hear a difference. And people pick the cheaper capacitors.
If it comes, if it comes to audio making changes, that are designed for taste, not for functionality, I always do it with substitution boxes, because I have to hear the switch or the change, it has to be almost instantaneous. Because if you don't, you will just load your mind up with bias. And you'll pick one based off of no real anything. And 99% of the time, you know, switching between two adjacent values on on Xbox. I mean, you never gonna know that you won't notice the difference between that even if you go up, you know, an order of magnitude A lot of times you don't even hear a difference between that.
Yeah, it kind of depends on what you're playing. So do you play through the the circuit while you're, while you're adjusting it in order to sort of see what it does? Or do you have a signal that you run through? How do you what do you use for the programming when you're doing that?
Yeah, it depends on the depends on what situation I'm going with most of the time it is playing through, and hearing and feeling the difference between it because a lot of times you'll make a change. And it doesn't sound any different. But I approach things different with with the instrument because it can change the feel, and not the tone. But maybe I'm getting a little bit into the weeds on that.
No, I it's really interesting to me, because I well, I got into electronics in the first place to work on my bass guitar. And now I'm at the closer to the end of my career, and I'm getting back into electronics to mess with my bass guitar. So in second childhood, as it were, I got a chance to try it again. But this time I know more. But um, what I'm finding out is there's a lot more to learn. And it's a big, complicated topic. It's very interesting. And yeah, hadn't really a lot of fun at Olympic working on the basis. It's great.
I wonder if this rolls down to we're gonna win more off topic, but rolls down to your senses are based off change and not absolute stuff. Like it's easier to see movements with your eyes than static things.
Yeah, like what Tim was just saying is he's basically made a differentiator, right? He's, he's putting in two settings and switching back and forth between them. It's kind of like a chopper, op amp, right? Where, where you, you have a reference, and then you have your signal, and you chop back and forth between them. And it's that delta that you see is much more sensitive than you know than it otherwise would be.
You have to be able to detect, identify, and even better yet be able to verbalize what changed to be able for it to actually manifest as a change itself. Anything else ends up being kind of voodoo in a way. So I'd say that when it comes to audio
Yeah, I need to get better at that. So what kind of speaker Do you listen to when you're just at your bench? And so how do you have your amplifier or whatever it is arranged Because I know with my workbench, I will have an amp next to it that I can hook up and hear things, but how do you do it?
Well, okay, so there's, there's a huge number of ways that I approach that. Just because if I'm developing a product for a specific situation, I tried to play it in that situation. Not men, when now we're really getting into weeds on things. I guess this is measurement in a way.
No, this is measurement. It's not with high accuracy test equipment. It is. It's the same thing, though. Yeah, it's your how you are testing it and how you approach it depends on the product and what you're trying to measure.
Right? Well, okay, so So let's, let's take great a great example speakers. Here, you know, speakers can be highly directional, very beamy. Straightforward. So if I'm trying to change, I'll have the speaker on the floor pointing at my ankles, and then I'll put it up on a desk pointing straight at my head, then I'll tip it at an angle, then I'll flip it backwards and hear it from behind, while I'm making the same change in all of those situations. And I'm making sure that I'm not just preferring how something is in one of those configurations, I'm making sure that I can hear that same change no matter what the orientation is. And things like with the speaker being so be me, if it's pointing at my ankles, I have I have a tendency to just crank treble, just so I can hear it five and a half feet high or six feet higher off the ground, than if it's on the bench pointing straight at my face, I have a huge tendency to to reduce treble. So, you know, if I'm, if I'm adjusting gain on a particular stage and an amplifier, and I'm doing a frequency dependent gain, that's where it really starts to matter. Because, you know, am I adding a ton of, of treble and distorting something in a in a unique way, that sounds great at my ankles, but we'll just, you know, saw your head off, if it's pointing at your face, you know, so it's a, that sort of goes back to what I was saying, like, earlier about, you know, when you're making a measurement, you got to know what you're shooting for. I've certainly made tons of mistakes by just putting something in the wrong situation. And, and going down a, you know, a five hour path of adjusting my circuitry. And, man, this is really fantastic. I love the sound when it's pointing out my ankles, and then I put in a just a slightly different configuration, and it's the worst thing I've ever heard. So yeah, it's so to answer that, it's it's super dependent on what I'm going for.
Well, it sounds like you get a lot of experience that is, is what you really need to do that. And so do you have that? Do you know of any good references for that? Or is is what you said written down anywhere?
I should write it down. So
your personal take on Yeah, it's actually quite interesting. I haven't heard it said quite like that before. What Who do you look for for audio? Reference materials, like I have books by an amplifier design, and some old books by Olson on audio. Do you have any favorites or any, anything you'd like?
I have one one favorite. And it's, it's well, Edie. He has multiple books, I think four different books on tube amplifier design, ranging from guitar amplifiers to Hi Fi. And one of the reasons why I really liked this gentleman's books is because they're, they're very, they're technical. They avoid descriptors as much as possible, and try to give specific reasoning as to why everything is happening. So it's, it's filled with equations, but you can ignore those if necessary. And and the underlying the underlying idea around the book is being able to wrangle the electronics to do what you want. So it's guiding towards a goal as opposed to saying, like, if you like this sound, do this. There's plenty of resource material out there that has that kind of feel to it. And I steer away from those because they're just, there's they're unbelievably subjective. So give me the equations, show me what's happening in my transfer curves or in the tube curves or or solid state curves, show me what's actually happening there. And then I can be the one who experiments on those edge cases, to find whatever characteristic I'm looking for. So that gentleman is Merlin Blencowe. And they he has just a handful of resource books about power supply design. Hi Fi and guitar amp stuff?
Well, I'll check it out. The one I like, just kind of introduction to design is the Marshall Leach. He was a professor at Georgia Tech, I think, Introduction to electro acoustics and audio amplifier design. And I think he, he used it as a textbook, I think. And it's got like full schematics in it with all the values and everything. And so you can actually build what's in there. And I've heard they work. I haven't built anything myself. But it's got, like you said, it's got a lot of equations in there. And one thing I like about it a lot is that the circuits are also relatively easy to simulate. If you put them into a simulator, they work. Unlike a lot of circuits that have some sort of unknown parameters that are, you're counting on such that if you put it in a simulator, if you didn't know to adjust this one little thing in this transistor or whatever, that it wouldn't actually do the right thing. These these I have actually simulated the circuits and they they work in fact, you can use the power amp as an op amp. If you need to design your own op amp sometime that's one a simulation model. Yeah. If you ever want a transistor based op amp so what are the what about tubes? Does this Merlin Blanco have tubes to cover tubes? Or just transistors? He
predominantly covers tubes.
Oh, okay. Cool. Yeah, my my classic for tube is is the radio Tron designers handbook. Jarrett. You ever see this thing?
I have two different versions of it. Oh,
okay. Yeah, there's a PDF of it. That's I
there's a there's an old textbook that I have, I probably told this story before. But there's an old textbook I have I love it. It was written in 4950. I can't remember what year but at the very beginning in the preface of the of the book, they're talking about basically what to expect and you know, this is the pinnacle of technology, these these newfangled vacuum tubes, and and they they start by discussing how accurate they have read the speed of light. And it's like, two, four or five digits, something like that. And it's and it's like this who robbed moment in the book was like, we know this very accurately.
Nice. So what's the name of that book? Do you remember?
That one? It's just an old textbook. So I'd have to I'd have to grab it. It's not a very academic name.
There's a very cool website. I think it's called two books.org. Is that right? No, that's not right. I have to find it. But anyway, there's a great website that has a whole bunch of PDFs of old electronics books. I'll have to see if I can find if I can find the. Yeah, yeah, there it is to book like the crack tube. books.org. Yeah. And so it has data sheets and lots of textbooks and other books, reference materials. Amazing, amazing. Reno.
You know, it's an old website, not by just how it looks, though. But it's, they're recommending me download Adobe Acrobat PDF reader.
Yeah. You can almost smell this website. You know, it's got that musty book. Musty news. You know, it's that book. Oh, that old book inviting? Yeah, it's like a used bookstore.
It was last updated August 24 2021.
Yeah, it's very active. The guy is incredibly productive and makes the most beautiful scans of things that people send to him. Or this just does a fantastic job. Yeah, it definitely focuses on the content and not the presentation. You know, it's all it's all what's inside the PDF is as well.
I prefer my websites to be like this because you can easily find the info. It's just the whole like, get Adobe Reader. Oh, yeah. Well, all browsers nowadays have that built in.
Under Construction sign. Yeah, yeah. Yeah, I definitely. Early days of the web, I was definitely one of the most definitely an early adopter. And so all that go by very cool
well, I guess we have anything else we want to kind of touched on, or
I think we've covered everything that that Tom has written down here. Yeah, I
think we've kind of gotten mostly over it. I'm sure we could go a lot more.
Well, if you want to. So I don't share this with too many people, but, but I'll share it today here. So there's a site called que tu rn f.com. That's kilowatt, two radio, Norway france.com. And that is a site that I built for my dad mostly. It's about his stuff. He was a radio amateur. And so you can you can check that out or put up a link to it. It has an old article that he wrote for Qs T. In 1965, and about his 15 minutes of fame as the first amateur to receive weather satellite photos.
Oh, I see. I've seen the PDF here. This is really cool. Wow, this is
yeah, I got permission to to, to put them up from QFC. They were very nice to me.
Very cool. So Tom, where can people ask you questions, if you want, we ask questions. Where can people find out more about you and what you do?
Well, the macro fed slack is actually a very reliable way to get a hold of me. And really great community. It's totally pristine and non polluted. It's one of the joys of the internet. And so thanks for hosting that and and keeping that good. Thank you, Tom. So I'm, I'm totally on there. And you can ask ask me anything, as many people do, and maybe I'll answer.
Okay, well, would you like to sign us out then?
That was the macro fab engineering podcast and I was your guest, Tom Anderson.
And we were your hosts Parker, Dolman
and Steven Craig.
Let everyone take it easy.