Pragmatic 72: So Much Noise

10 April, 2016


Noise as a term is widely misused and misunderstood. Carmen joins John to sort out the signal from the noise.

Transcript available
Welcome to Pragmatic. [Music] Pragmatic is a discussion show contemplating the practical application of technology. By exploring the real-world trade-offs, we look at how great ideas are transformed into products and services that can change our lives. Nothing is as simple as it seems. Pragmatic is part of the Engineered Network. to support our shows, including this one, head over to our Patreon page. And for other great shows, visit today. I'm your host, John Chidjy. And today I'm joined by Carmen Parisi. How's it going, Carmen? Pretty good, John. Glad to be back on the show. That's good to have you back. Yeah, it's been a while. We've both been very... Glad we sorted out my payment, you know, thing. I wanted a raise, damn it. Oh, yeah. Well, this is what happens when you get pushy, isn't it? Geez, I don't know. Yes, yes it is. I don't know. You almost blacklisted me. I don't know. It's terrible. But no, actually, truth be told, we've both been really busy at work, actually, because, yes, work, work, work and all that. And I'm about to be disappearing off on holidays. But anyway, that's all right. So I, because of your experience with electronics and your background, I thought a perfect topic for us to talk about would be noise. Oh, wonderful. Yeah. My statistics math isn't really up to snuff since I left school, so let's see what I can do. Oh, that's OK. It's all good. I mean, the thing about noise though is, well, you can't escape it. You've got to start with that. You can't escape noise, but it follows you from one room to the next or something like that. But anyway, how about you start off with how would you define noise? Sure. So noise by definition, it has to be from a stochastic or random process. So unfortunately, it's become a catch-all term nowadays, but although you're measuring you know, signals, clocked signals where they shouldn't be, or you're hearing a repetitive audio audible sound. Those aren't noise because they're not random. You can trace them back to a source and it's just coupling or radiated emissions, something like that. Yeah, exactly. I find it very frustrating how a lot of people will say things like, yes, we've got a 50 hertz noise on our audio or something like that. It's like, well, technically that's not a noise. I mean, it is a noise, but it is not noise. OK, it's interference technically, but it's a sound you do not intend, but it has a regular. It has a regular source, but anyhow. All right, so I'm guilty of it, too, from time to time, I'll look at one of our feedback traces and say, oh, that was a poor layout. There's a lot of phase noise on there when really I know exactly where it came from, because it came from our phase node. Yeah, I know. I understand. I guess it's kind of frustrating. And when I do that, I slap myself as well. But, you know, as long as you get the message to who it needs to be. Yeah. Well, this is the thing. And we need to receive it. Well, this is it. OK, so this it comes back to we're trying to get a message from point A to point B, whatever those points may be, whatever that message might be and so on and so forth, whatever the transmission medium might be. And I always love the expression pulling a signal out of the noise, even though technically that's like, well, in many respects, I don't know. Is that ridiculous? It kind of is, I think, because you don't really pull signals out of it. I conjure up images of someone like with gloves trying to pull a signal out. It's like there's a bit of noise and they're trying to pull it out. It's like, hmm, okay. I have weird images in my head, clearly. Anyhow. We won't touch on that. The psychology episodes for later in the year. Oh, God, that would be a very scary episode for us. I'm just saying. I'm sure it was. I could record from the couch next to me. I just lay down like we're in the psychologist's office. Now, where did you first hear the noise, Carmen? Anyhow. - It all started when I was young. - Oh dear. Okay, I'm gonna stop you right there. Okay, so here's the thing. Signals and noise. So of course, people will hear about this thing called signal to noise ratio. And ultimately, I guess the thing with signal to noise is it's a double, it's a ratio, it says what it is on the box. It's the ratio of the maximum signal intensity to your noise intensity. And ultimately, if you wanna improve getting your chance of a message getting from point A to point B, you got two options. You can improve the sensitivity of your receiver, or you can increase the amount of power in the signal that you're transmitting. So ultimately you can increase your signal or you can, well, essentially make it more sensitive and less noisy at the receiver end. So I guess when I say signal, I mean, that's the information we're trying to transmit. It doesn't matter how it's transmitted, whether it's narrow band, broadband, you know, spread spectrum doesn't really matter. You know, just a signal. For the purposes of this episode, we won't go any further, but let's just say that's what it is. But receiver sensitivity comes back to how much you can amplify that signal and how much noise it amplifies in the process, but also how much noise it introduces in that process. And that's a part that I guess I want to explore more, is the fact that noise is the part of that that you can't quite escape. So the way I think about noise is people say, well, you know, if only we had like a lower noise amplifier, then we'd be able to get away from, you know, we can amplify this signal, we can amplify this audio, we can amplify this radio signal, whatever. If only we had a more selective, more sensitive receiver and we can amplify it and separate that from the noise. And I guess the thing is you can't really escape, you can't escape noise because noise is sort of the, to me, it's the electrical equivalent of friction. You'll always have friction. you know, of some kind, like mechanical friction, either aerodynamic friction. I guess you call that drag or I mean, ultimately, though, because air acts as a fluid, then that's actually a fluidic friction. Then you've got kinetic or otherwise known as dynamic friction, sliding, rolling, static, different kinds, but you've always got friction. Ultimately, you can never overcome it. There's no such thing as a frictionless anything, even in space. There is in physics one classes. Oh, that's just not fair, though. That's not ideal. No, there's technically even friction in space. There's still a molecule every cubic metre or whatever the heck it is. It's not such a thing as a perfect vacuum. So, technically, you've always got some kind of you've got some kind of resistance. So, you can't get away from that even in space. Just- Those hydrogen atoms really slow our space shuttles down. Damn straight. When I'm going at warp speed, that hydrogen really pisses me off. Anyhow- You can't, you can never clean it off the windshield. That's what's causing all those marks. Now I get it. Okay, so actually funnily enough though, we're just talking about fluidic friction. If you want to, there's a great podcast that I've heard of, it's called Nutrium and it's actually on this network. So, if- No way. Yeah, I know. How about that? Wait, there's more. There is always more on the Engineer Network. And if you want to learn more about Bernoulli's equation and all that other great fluid dynamic stuff, there's a couple of really great episodes of Nutrium about that. So, there'll be links in the show notes if you're interested in that sort of thing. Very cool stuff. Anyway, so, but as I said before, noise and interference, two very different things. So I guess there's many different kinds of interference. And I guess the one that comes to mind is it's like, it's an adjacent signal. Technically, it's not noise, but it's interfering with the signal that you're trying to extract. And it takes me back to, or do you remember the Artemis P cell? Steve Pellman's, well, I don't recall. Oh, yes, I remember that episode. You did it a long time ago. It was about two years ago. Yeah, that's right. And still, it was one of my favorite pragmatic episodes. Oh, cool. Thank you. Yeah. Well, that was episode 16, I think it was. Yes, that's right. One man's hopes and dreams of an RF bubble. And that's still a hope and a dream, by the way, at least in the real world. I think it works OK in a lab. But anyway, so I guess the thing that comes to mind when we think about interference and so on is that there's a lot of misinformation going around about P-cell. And again, like I said, where that works in the real world, you know, I'll be impressed when they can make it work in the real world. But anyway, and there were expressions that were used in some of the reporting at the time, tech reporting, like it uses interference to its advantage. But then you'd read other articles that would mutate that to it uses noise to its advantage. And it's like, well, no, that's not possible because that's not what they're doing. But anyway, either way, whether it does or there's open to debate, I guess maybe I'll revisit P-Cell one day if they ever get it to work in the real world and maybe I should just leave that there. So rather than me keep talking for a minute, what about, let's just focus on noise and solely noise and only noise just for the moment. And I think that they refer to noise in different colours. Can you tell me a little bit more about that? Yes. So just because we say noise has to be random, it does not mean that it all has to be the uniform white noise static on the TV, you know, stereotypical image that gets conjured up when you think about noise. There are colors and it refers to how they react over frequency and some of the most common colors of noise are white noise, which is flat when you look across a frequency band. There's pink noise, which follows a 1/f characteristic and decreases with frequency. brown or red noise follows one over F squared so it drops off even faster over frequency and then opposite pink and brown noise there's blue and violet. Blue and violet. You said violent? Violent noise. Violent noise. Oh no. Yes it wrecks your receiver. That's like really scary man I don't want any of that stuff. No no you really got to filter that out and those two colours increase with respect to F and F squared. And then there's some other random noise like grey noise and I can't remember. There's oodles of noise. Noise, noise. A whole rainbow of noise. It's a wide spectrum of noise. Oh, dear me. Yes, violent noises. That's a good one. I remember that. Anyway, I'm not going to let you live that one down either. OK, cool. That's fine. It's on our greatest hits track. Oh, that's terrible. Anyway, OK, so sorry, keep going. The noise we're going to be talking about today is AWGN, which stands for Additive White Gaussian Noise. And breaking that down into digestible chunks, clearly I have a voice for radio today. Additive means it's on top of any noise intrinsic to the system, whether it's a control system, electronic circuit, or information system. You have your clock noise, your jitter, your whatever we've been talking about already, and then the noise that is just there. It's a very meta concept. Far out man. White, which means it's uniform across the frequency band of interest and Gaussian, so the noise follows a normal distribution when you look at it in the time domain and it has an average value of zero. So in electronics terms it means there's no DC component to the noise signal. Yeah, Gaussian noise, it took me a while to get my head around how that was possible but But yeah, I finally got my head around that after a few years playing with noise generators and such. But yeah, it is fascinating. I love the term AWGN because it's just, it covers a multitude of issues relating to noise once you get your head around that, I think you can get your head around noise. - Yeah, I just barely scratched the surface of trying to model noise and go over that in school. it was it got pretty trippy pretty fast. Oh yeah, but it's all good. So, oh dear. There was something else that you've mentioned that that bears mentioned something that NASA launched. Oh yes, the Cosmic Background Explorer or COBE mission. It was launched back in November of 1989 and the purpose of this mission was to map the fluctuations of the background radiation across the known skies and help improve our understanding of the early universe. And what it discovered was that a background radiation is real and that the white Gaussian noise that we're going to be talking about is in the very fabric of the universe itself and is actually left over from the Big Bang. Why it's left over from the Big Bang? You're going to have to ask someone smarter than me or who has more physics knowledge, but it's everywhere. and the static you see on a TV, an old TV or the hiss of an AM radio, that's actually you're listening to remnants of the Big Bang itself, which is pretty cool. Yeah, exactly. I mean, it's fascinating. They made a whole TV show about it, apparently, the Big Bang theory and background radiation or something. I think it's a physics show. Yeah. I have to watch that sometime. I don't know what it has to do with eating, you know, what the Big Bang has to do with eating Chinese food in an apartment. But well, it's closely related. It's closely related, you know. No, I can't make that work anyhow. So, yes, we are not comedians. No, no, we're engineers that we had the comedian beaten out of us at uni. I speak for myself anyhow, possibly. I got I got more wise than comedian. Fine. That's OK. Partial credit anyhow. So, yes, the thing about background radiation, if I remember correctly, is that there are actually, there are hot spots when they map the galaxy looking at the background radiation. I thought there were warmer spots than others, but there was always a minimum level of background radiation, if I remember correctly. Yeah, it's down at minus two or minus three Kelvin. Yeah. Or sorry, two or three Kelvin. There's no negative Kelvin. One would hope not. Otherwise, the definition of zero Kelvin being zero molecular movement. Yeah, so negative molecular movement would be like antimatter or something I guess. I'm just trying to think how even that would work. That wouldn't work either. I don't know. I think we blew up the universe with negative Kelvin. That's what happened. That's where the dark matter lives man. It's true in the negative Kelvin region. That's what it is. There you go. Now we've just accounted for all the missing matter in the galaxy. Nice. There we go. So yeah, there were fluctuations but they were small fluctuations given you're going across the size of the universe and there's a two Kelvin difference, that's pretty good. Oh, yeah, absolutely. So, it's interesting because the noise level that you mentioned then, was also expressed, is expressed generally as thermal. So, you get like a thermal noise levels is an indication, noise temperature, actually, I think is its technical term. And so, when we look around, we, especially in radio astronomy, they'll talk about that. And it becomes a method of measurement of thermal noise in low noise amplifiers, particularly used in astronomy, which is really cool, I think. So, might as well talk about that now. So, yeah. So, in electronics, we talk about noise temperature as an expression of the level of available noise power introduced by a component or a source and methods of reducing the amount of thermal noise in low noise amplifiers these days usually involves cooling the LNA silicon to an extremely low temperature. And originally, if I remember correctly, there was liquid nitrogen. And liquid nitrogen hangs around the 63 Kelvin mark, I think is its transition temperature, I think from memory, which is minus 210 degrees Celsius, that's minus 346 degrees Fahrenheit. So, that's pretty cold. But the great thing about liquid nitrogen though is it's cheap. It's right, it's quite easy to make liquid nitrogen. These days, it's made in bulk because I went to doctor's surgery and one of my children had a small wart on their thumb, I think it was, or one of their fingers anyway. And they said, oh, we'll just freeze that off. And I'm looking at it and I'm like, freeze it off. How are you going to do that exactly? Like get an ice cube? Silly me, had no idea that they just had a canister of liquid nitrogen out the back. And I'm like, OK, cool. When I was a kid, we did not have this at the back of the doctor's surgery. So, okay. And that's not a hospital. Doctors are high tech these days. No, but it's just a doctor's surgery, you know, like a standalone building in the middle of the street. Family practice. Yeah, that kind of thing. Yeah. Not a hospital, not a massive, you know, 10 story building with ventilators and resuscitators and, you know, backup generators. It's just like your average run of the mill, you know, family practice. And there is a big vat of liquid nitrogen. And I'm looking at this and I'm like, oh, cool. that's also dangerous. But anyway, so they take a bit out and they put some in a small canister, they put the lid on it, and it's got a nozzle on the end. And all it is, it's basically it's a flow control nozzle. And as you press down on the button, it just opens it up. And as it reaches the end, of course, because the pressure has been released, it comes out in a jet of vapor, but it's highly focused. And it's very, very cold. Not minus 210 degrees Celsius cold, but certainly cold enough and they go freeze the water off and drops off and it all heals and it's all good. So there you go. Anyway, that had nothing to do with liquid with noise amplifiers, but never mind. It's noise in the show. You got to extract the nuggets of useful information from our babbling. Oh, that's so good. Okay, cool. Very good. So but anyway, liquid nitrogen was used for ages because it was cheap, you know, and that's fine. I totally get that. But of course, the cold, How coal, how low can you go? Well, go down to liquid helium. That's pretty much as cold as you're going to get in terms of like, you know, because liquid hydrogen is well, you know, rocket fuel essentially. So, kind of don't think that's safe. I don't want my doctor playing around with rocket fuel. Let me just let me just burn that off there. Oh, sorry, I burnt your whole hand off. Sorry about that. And the rest of the office behind me. Yeah, sorry about the hole in the wall. Yeah. Anyway, but we got rid of that. Well, anyhow, so liquid hydride, liquid helium, I almost said hydrogen. So liquid helium, it goes down to a little bit of a frosty 4.2 degrees Kelvin. So that's kind of cold. Minus 269 degrees Celsius, minus 452.2 degrees Fahrenheit. Oh, dear. And thank you for correcting those show notes as we went. Much appreciated. Yes. You're welcome. Dear listener. Yes, I actually had liquid hydrogen and he corrected the symbol on the periodic table to helium. Thank you, Colin. So, yes, anyway, and there's a great link. Well, okay, I think it's a great link in the show notes. It's a big PDF. So brace yourself, you're going to have a look at it. And there's a chart in there that shows the thermal noise temperature relative to frequency. And it's got two plot groups, one at 300 Kelvin and the other at 80 Kelvin. And that just shows the relative difference between, well, it's not quite, it's not exactly room temperature, but call it room temperature if you'd like. And of course, 80 degrees at around about around about liquid nitrogen temperatures. And at 20 gigahertz for this LNA, you got about 100K of difference in your noise floor. And by the time you get up to 100 gigahertz, the improvement's enormous. It's like 350K reduction in noise temperature. Yeah. And that's just going to liquid nitrogen. So, clearly the contribution from that, from the thermal noise through the silicon is reduced significantly at lower temperatures, because of course, with less- Is it? Yeah, with less molecular movement, you know, you've got less noise being generated. It's just, yeah, fantastic. So, they had like 40 dB of signal to noise ratio and then they dumped some liquid nitrogen on there and all of a sudden it shot up to 390. - Yeah, it basically gets to the point where there's another chart in there that's really interesting as well. So it shows the signals just above galactic noise. And prior to them using liquid nitrogen, there was simply no way for them to extract that information from this, extract that from the noise, because the noise of the LNA was drowning out the signals that they were trying to receive in the radio telescope. So it was an enormous improvement. So it's- - That's really cool. - Oh, it's very cool. Oh yeah, very cool. I could have used this on my senior design projects. (laughing) - Yeah, very cool, see, anyway, all right. So yeah, I always found that fascinating. And actually that movie "Contact" with, geez, what's her name? - Jodie Foster. - Jodie Foster, thank you. Yep, amazing, amazing actress. And she did a fantastic job in that movie. And it was in there, one of the things about her story was that, well, the story of the scientist that she played was that she helped with the evolution or discovery of making the radio telescopes more sensitive. And that was quoted as one of her achievements was cooling them with liquid nitrogen. But anyway, if I remember correctly. It's a good movie, check it out if you haven't seen it. Right, so moving on to audio applications. Now I know that I talked about audio noise way back. That was in episode eight, way, way back. And that was more with respect to loud noises or maximum sound pressure levels. But I guess now we're talking about noise. I just want to flip that on its head. And I guess the other reason is that I've been delving more into audio production these days with podcasting and microphones and amplifiers. So, I think it's interesting to explore just briefly what we're talking about noise. So I guess fundamentally, if you mess up levels and so on in an amplifier and the noise and the... you recorded... How do I start this? When I'm speaking into a microphone and the microphone is amplifying a signal from my voice and sound pressure waves turning it into an electrical signal, if that signal exceeds the maximum limit of the amplifier, then it has this effect called clipping and that clipping creates a distortion. Now, technically, that distortion isn't noise. It's distortion. That's what it is. And if you've got an amplifier, the amplifier is plugged into the wall socket, power point, whatever you want to call it, and it's got 50 hertz or 60 hertz AC, that technically is not noise, that's interference because you're picking up that signal and it's interfering with the audio that you're trying to record. And if you're coupling, let's say you've got a switch mode power converter, a point of use power supply or a switch mode power supply powering your laptop or your computer or the amplifier, whatever you're using, and it's getting some coupling and you're coupling some of that rectifier noise that's coming directly over the power lines even. Well, coupling is again, technically it's an interference. So, it's not the noise that we're talking about. And so, I guess the problem I have is when people say, oh, you got noise in your audio. It's like, well, you have a noise in your audio. Yes, you do. But technically, it's not noise. So, yeah, that 50 or 60 AC hum, you can isolate. If you isolate your amplifier better, and this is something I was hoping you could talk a little bit about, because I know you've built amplifiers and so on previously. Correct me if I'm role? Yes. I played around with them a little bit, but I work in power supplies, so I know a little bit about decoupling and all that good stuff. Well, so this is the thing that's always bugged me, because when I was doing- When I was in amateur radio, there were a lot of people that would say, well, if you don't operate off of a battery. And the early radios were originally designed to be battery powered. So, you had a lead acid battery at 12, 13.8 volts floating or 24 volts floating, whatever it was. And it was the smoothest signal you could possibly hope for. It's a battery, you know. And once you went to a base station setup, you had to go with a power supply and you had linear power supplies that would essentially step down the AC, run it through a transformer, run it through a rectifier. And then normally you would then essentially just you'd regulate that. So you'd regulate that and you'd have a very hot diode in there, ultimately, and essentially try and smooth the top off. With electronic controls, sometimes they had a bunch of smoothing circuits and it was all relatively clean if you had a good design. But then there were people that were saying, well, let's go switch mode, you know, because switch mode power supplies, I can chop it all up precisely. I can have PWM in there and I can just I can chop it up into a thousand pieces. We'll go over a 600 volt DC bus, and then we'll chop and change it back again into whatever and step it up and step it down as we need to. And those sorts of switch mode supplies, boost converters, buck converters, all that other good stuff. And just forget the transformer. If we use a transformer, it's more for smoothing purposes than it is for actual switching purposes, depending upon the kind of converter that you've got. And, but all that stuff, the initial designs were so terribly, they produce so much interference. Yeah, we've come a long way in quieting down switch mode power circuits. Oh, yeah. So, I mean, yeah, there's plenty of techniques depending on exactly which topology you're using. You can slow down the high speed nodes that, you know, radiate noise everywhere. And, you know, the output's really easy to filter. They always say you get that ripple in there. Well, you can beat that down if you just throw giant capacitors on there. Big tantalums, right? It has its own... What was that? Big tantalums, huh? Yes, yes. Tantalums or even ceramics do pretty well nowadays. You can get a... I got some 47 microfarad caps in a 603 package on my bench at work. Which is very small for those of you. Yeah, 603 package is 6 mils by 3 mils and 47 mics is a pretty large value to get in such a small package. Yeah, that is. That's a ceramic. Yes. Nice. Man. It's without turning this into a capacitor episode. Yeah, it's an X5R dielectric. Nice. Where were they? So not top of the line, but still pretty good. Where were those things when I was doing this? Works for my work. I haven't. See, the thing is, it almost makes it too easy to use those. You got to find customers who would pay to have that capacitor because it's still pretty new and, you know, relatively hard to find. Yeah, the good engineers design well with the cheap caps. Yeah. Well, it becomes a trade off, right? Because your trade off is how much cost you want to put in to essentially simplify your design. But, you know, straight up, you're going to have a far more expensive end. I mean, how many suppliers actually supply them? What are their stock levels? And, you know, there's a whole bunch of other variables. And, you know, the end application makes a big difference, too. If it's if customers build a standard laptop, you know, they have a little bit more space. But some of these new tablet designs I've helped work on, the motherboard is just insanely small and they'll pay because there's just no room to put more caps on. Yeah, well, that's exactly right. People don't realize that they think that we can just keep miniaturizing and making this stuff smaller and smaller. It's like, well, actually, there's some you do hit the wall at some point. You need some new technology like some of these high capacity capacitors that are actually made of something other than tantalum. I mean, tantalum is good and bad, right? I mean, when I was doing at Nortel, we had no end of trouble with tantalum caps and we reached the point when our designs where we banned them for it. They had to improve those too. I don't have all my capacitor notes on me, but it was because there was a lot of oxygen in the design of the chemicals for the capacitor. So, when they would blow up, they would burn, but they've moved away from that. Yeah, when they burnt, they burnt good. I mean, we had, I mean, circuit boards, literally, I think we had like a six ply board, a six layer board that was, that literally had a hole burnt through it. Not like, not a huge gaping hole, but you could see light through it, you know, from a tantalum that lost its lid. It's, and we had years of statistical data saying, you know, this tantalum's a bad, stop using them. I know that they've come a long way. That was 90, geez, that was 99, 2000, I think it was. So it was 15 years ago. So I know things have changed for tantalums. - Yeah, they've come a long way. They got no burn tantalums that fell open instead of short. - Yeah, that's a step forward. It's become the capacitor episode, making noise about. - Yes, it has, more noise. - That's terrible. Anyway, all right, good. So, all right, anyway, so where were we? We're talking about amplifiers and AC hums and all that sort of stuff, okay. - Yes, yes. The main point of all that ranting and side tracking was if you lay out your circuit properly, if you use some low noise design techniques, if you really know your power supply stuff, you can get rid of the 60 Hertz hum. - Yeah, exactly. But there really is no excuse. And honestly, it's interesting because people say, oh, well, I'm plugging in this amplifier and this amplifier is plugged into USB. Oh, but how come it's got noise on it? Because USB is just like 5 volt DC and isn't it like, well, yeah, but what's powering the 5 volts? It's probably a switch mode supply and it's probably not the cleanest and- It's definitely a switch mode supply. Yeah. And it's like, well, then, but it's battery powered. Yeah. What voltage is the battery and what voltage is- Okay, so it's 5 volts on USB and your batteries. Oh, look at that. It's 12 volt battery. So, okay, how do you get the 5 volts? And then the penny drops. - If we want to do an episode on battery chargers, they're so cool. - They are actually, they are very cool, especially the lithium ones, but anyway, all right. Another topic for another day come. So if on the amplifiers, I guess, the other thing about amplifiers, I was saying before about clipping, just a little bit more about that, it's you want to amplify a signal in its linear region, obviously, because if you're amplifying in a non-linear region of the amplifier, then you're going to get distortion. And that distortion is going to be interference over the final signal that you're trying to amplify. And the flat topping is the worst, or clipping is the worst, because ultimately, as Mr. Fourier said, and we actually talked about this relatively recently on episode 64 about software radios, and if you wanna learn more about Fourier transforms and so on, but clipping causes wide spectrum interference because square signals equal bad, and that's bad. So anyway, if you wanna know more about that, then have a listen to that episode. So lots of crossovers in other episodes here, bit still. But in the end, I guess for most spoken audio anyway, once you consider the noise that's introduced by transmission and then recreation through speakers or headphones, I guess there's no reason to get liquid nitrogen cooled preamps for your mic, but I'm sure there's someone who's going to build one and they'll sell it because gold-plated stuff sells too, right? On the same kind of, you know, snake oil idea that it actually makes a difference. But anyhow, never mind that. Can you imagine? I have a liquid nitrogen powered amplifier for my mic. That's what I'm speaking into right now. That's why you sound so good. Anyway. Exactly. Yeah. I've also lined the walls of my room with gold, so I get proper sound isolation and oh, man. Maximum interference. I'm lacking- I don't have enough gold in my life. Anyway, that's all right. That works on two levels. Okay. Human hearing is this- it's just- it's very imperfect. It's one of those things though, that's kind of cool the way it works and the decibel scale and everything. Again, we covered this way back in episode eight, but zero dB on the sound pressure level scales calibrated to the minimum amount of sound pressure that a human can hear. And that's at one kilohertz, because why not? 1000 hertz, I guess, which is roughly not quite the middle of the human hearing range, but certainly it's a good enough point as any. And that pressure is two by ten to the minus five pascals of pressure. And that's measured at one atmosphere. So. Seems like an incredibly small amount of pressure. Yeah. And the human ear can detect it. That's the threshold of human hearing. So. How far away is that? That's like beyond a whisper. Yeah, I know that's at the ear level. That's actually at the ear entrance to the ear canal. That's how much closer you got to be to hear that, because about 10 dB more than that is rustling leaves blowing in the ground. So, just quite rustling of the leaves. So, it's very, very soft. But the funny thing is, even at those sorts of levels, the limits of human hearing obviously being what they are, there reaches a point very quickly, there's no additional benefit of putting more and more money into better and better headphones or amplifiers or microphones. Again, not just the recording side of it, but the recreation for the listening side of it as well, because you never improve beyond a certain point, because the average person will not be able to tell the difference. And if you've got tinnitus like I've got it, then it's not even that worth it, that amount of money and effort, because to overcome the sound of the tinnitus, the signal has to be at a certain level. And I'd never hear the noise anyway. I just think it was the tinnitus and oh, dear. Well, I'm actually undergoing an operation next month to get liquid nitrogen cooled bionic ears, so none of this applies to me. Nice. Well, let's just hope the liquid nitrogen doesn't leak, because that could be really bad. But anyhow. Brain freeze. Yeah, just when you thought slurpees are the only things that could give you brain freezes. I haven't had a slurpee in a while. Yeah, it's the wrong time of year for you, but anyway. Okay, so actually on the tinnitus front, I found something recently that blood sugar level can actually make tinnitus worse. So, there you go. Interesting. Yeah, I found out that recently, like recently in the last couple of months. So, there you go. Like too high or too low? Too much blood sugar. So, if your blood sugar is too high, then your tinnitus gets worse. So, I've discovered. So, I'm now using my tinnitus level as a way of gauging if I've had too much sugar. Yeah, surprisingly accurate. But anyway, never mind that. So, if you have a slurpee which is full of sugar, your tinnitus will get worse, but you're colder so it should make it better. Yeah, now you're trying to get fancy. Slurpees are even, they're neutral, they cancel out. If only it were like that. No, actually what I've done on the subject of slurpees is I've actually switched to the 1% sugar ones, so they're really low. Oh, okay. Yeah. they 7-Eleven, we actually have 7-Eleven over here these days. Didn't when I was a kid, but we do now. And they'll have a Slurpee there that's always the the low calorie, low, low, whatever you want to call it version. So, that's always the one I get if I get one. I think I've had maybe three or four this whole summer. We're almost at the end of summer now, so, you know. And yes, they're not so bad. But anyway, there you go. I actually don't have too much more to say actually about the subject, I guess, other than the fact that you can't escape noise. It's just a fact of the fact, molecular noise, you'll always get it. You can't fight it. You can't escape it. It'll always be there. And unless you want to go on liquid cool stuff with liquid nitrogen, liquid helium, you'll never really reduce it significantly. Different technologies, different amplifiers are better at not producing as much. but honestly, just be aware also that the trade-offs, I mean, it made a big difference in radio astronomy because they're trying to pull signals down from the galactic noise level and the amplifiers were just too loud. Now, they're not because they cool them, but for audio systems, it makes no very little difference. You spend like $100 on a microphone, it's probably about all you really need to spend to get a decent noise performance out of it. It's amplifier, maybe a couple of hundred bucks, headphones, you know, 30, 40 bucks, you know, don't have to spend stupid amounts of money. And yeah, I don't know what else to say, really. Yeah, I don't know. Low noise design is pretty interesting. My my senior design in school was a low noise system. We were measuring where you designing a low noise system to measure noise on transistors. Cool. Yeah. And it was it was driving the senior design people nuts because our project did not fit into the rubric that they thought it should follow. You know, why didn't you choose this amplifier? The noise is too high. But what about the other features? It doesn't matter. So did they make a lot of noise about it? Yes, they did. Yeah, we actually met the noise specs, but the power specs were not not met at all. We were sucking through nine volt batteries like you wouldn't believe. Well, that's OK. You just always plug it into a power supply on the wall. Hang on. That was given the lab setup we had to install this in that was easier said than done. Oh, man. Yeah, the 9 volts lasted just long enough to take like one measurement and then you had to replace the 9 volts. It's very expensive. That's still pretty cool though. One of the things that a friend of mine did because he had noisy, it was a noisy linear power supply for his base station set up years ago, ago is he actually had a solar power solar panel for charging a battery bank outside his window. And he also had the mains charger as well. So, he would never operate off the mains, he would only operate off the battery, but he would use the mains to charge the battery. When he wanted to use his radio, he would turn off the charger, and then he would simply operate the radio from the batteries, which would also sometimes get a trickle charge from solar panels. So, that was the way he got the cleanest possible signal. Mind you, he could have just put a better quality power supply, but that was the other option. Yeah, the engineering to work around the poor power supply is way better. Well, I mean, it had the virtue of being very successful. It worked quite well, but you know, and he had a very clean signal. Good for him. But yeah, I just- We were trying to reduce our power consumption and, you know, battery draw by- You know, we couldn't really put anything to sleep because we were measuring one over f noise, and we said it falls off with 1 over the frequency, roughly. And the corner frequency, where the 1 over f noise meets thermal noise for our transistors, was at like 100 hertz. So to get a good signal, a good characterization of this noise, you had to measure down at like millahertz and just long time scales. So you couldn't put anything to sleep, or you'd ruin the measurement. So you almost had to be active the whole time. Oh, cool. Excellent. And our digital stuff that we did shut down to not, you know, couple in on the measurement that didn't have to do anything anyways, that was like nothing on our current draw. That's pretty cool. I wish I had it. That was interesting. Yeah. I sort of, yeah, we did. I did some noise factor measurements way, way back on an experimental board that never saw the light of day. but it was interesting stuff and dear. But anyway, enough about that. I didn't have too much else to add, so we might wrap it up there, I think, for this one. OK. Yeah. So before we throw in more noise to the podcast. Yeah, yeah, I think so. I think we've I think we've well and truly beaten that one, that pun to death, too. But that's OK. We're just going to keep running with it. If you want to talk more about this, you can reach me on Twitter at John Chidjy or you can follow at Pragmatic Show to specifically see show announcements and other related stuff. As you know, Pragmatic is part of the Engineered Network and also has an account @engineered_net that has show announcements about the network and all the shows and you can check them all out at today. People are really loving Causality. That's a solo podcast, if you haven't heard already, and that I do, and it looks at cause and effect of major events in history. So if you're a fan of this show, you may like it too. So be sure to check it out. If you'd like to get in touch with Carmen, what's the best way for them to get in touch with you, mate? Best way to get in touch with me is either on Twitter, I'm @FakeEEQuips, or you can go to my other podcast website, The Engineering Commons, and there's a contact form there as well. - Fantastic, and if you're not subscribed to The Engineering Commons, you definitely should. If you're an engineer, you will love it, there is no question. But anyway, if you'd like to send any feedback about the show or the network, please use the feedback form on the website, and that's where you'll also find the show notes for this episode. So if you are enjoying Pragmatic you want to support the show, you can like one of our backers, Chris Stone. He and many others are patrons of the show via Patreon, and you can find that at or one word. So if you'd like to contribute something, anything at all, it's all very much appreciated. So a special thank you to our patrons and a big thank you to everyone for listening. And as always, thank you, Carmen. All right. Thank you, John. Always a pleasure doing the show with you. Thanks, mate. [MUSIC PLAYING] [Music] [Music] (upbeat music) (upbeat music) (upbeat music) (upbeat music) (upbeat music) [MUSIC] [MUSIC PLAYING] [Music] (thunder crashing)
Duration 42 minutes and 41 seconds Direct Download

Show Notes

Related Episodes:

Neutrium Episodes:

Miscellaneous Links:

Episode Gold Producer: 'r'.
Episode Silver Producer: Chris Stone.
Premium supporters have access to ad-free, early released episodes with a full back-catalogues of previous episodes


Carmen Parisi

Carmen Parisi

Carmen is an Electrical Engineer working as an Application Engineer in analogue electronics and has a blog Fake EE Quips that he occasionally posts to. Carmen is also a co-host on The Engineering Commons podcast.

John Chidgey

John Chidgey

John is an Electrical, Instrumentation and Control Systems Engineer, software developer, podcaster, vocal actor and runs TechDistortion and the Engineered Network. John is a Chartered Professional Engineer in both Electrical Engineering and Information, Telecommunications and Electronics Engineering (ITEE) and a semi-regular conference speaker.

John has produced and appeared on many podcasts including Pragmatic and Causality and is available for hire for Vocal Acting or advertising. He has experience and interest in HMI Design, Alarm Management, Cyber-security and Root Cause Analysis.

You can find him on the Fediverse and on Twitter.