Linus Tech Tips Pushes Audio Grade Capacitors

[free_electron]

Explain me this : the entire data tract in a computer is digital. Even the amplifier is a PCM type class D or T.
What signal path capacitors do you need ?

Audio leaving the computer is over USB , wifi , ethernet or SPDIF ... or other digital formats.

[Rerouter]

The AC coupling capacitors to the outputs and line in / mic, and front panel out / mic

People use headphones and external speakers, even computer monitors with speakers give a 3.5mm jack input option.

[floobydust]

Electrolytic capacitors can cause high THD due to dielectric absorption, this has been known for decades as well as the increasing ESR above ~10kHz older parts have. It depends on the circuit, impedances and size of the cap whether the distortion is significant. It can be highly audible.

I didn't care much for boutique audiophool parts, then I see Nichicon's Muse capacitors with "silk" dielectric and realized science crossed into quackery? Nichicon makes at least 10 different audio electrolytics with nothing but psycho-acoustic terminology, 85C only rated and no data on DA.

Douglas Self and others measured the distortion caused by electrolytic capacitors, so there is some science that can be done.

[wraper]

Audio leaving the computer is over USB , wifi , ethernet or SPDIF ... or other digital formats.

In most cases none of the listed. Simply analog output to either speakers with amplifier or headphones.

[0culus]

Surprises exactly nobody - the guy is pretty clueless and not only about capacitors. His video about "fixing" videocards with a heatgun is another such example - and earned him a pretty savage response from Louis Rossmann.

If anyone is relying on Linus for their technical information, I am pretty sorry for them.

Man, I missed that Rossman video, gonna have to go find it.  I don't always agree with him, but I definitely appreciate his no BS attitude.

As to Linus, I guess he must appeal to people who don't know any better judging on his channel size. I find him exceptionally annoying, and the punchable face he puts in all his video thumbnails is the cherry on top. 

[EEVblog]

AVX have "audio grade" capacitors, and they watch my videos. I'm going ask them to prove their marketing claims, and in what way their "audio grade" are manufactured differently from their regular capacitors of the exact same spec. And how they actually quantify "deeper bass" and all those wank words. Let's see if I get a response 

[EEVblog]

As to Linus, I guess he must appeal to people who don't know any better judging on his channel size.

When your channel gets to that size you get dozens of offers a week for paid promotions. He has to take these in order to keep his large staff media empire going. It's a vicious cycle.
I get a couple of offers week but turn them all down.

[GeorgeOfTheJungle]

Don't you know that all this audiophile grade gear sounds even better yet if fed with clean, renewable power electrons only?

[EEVblog]

As to Linus, I guess he must appeal to people who don't know any better judging on his channel size.

When your channel gets to that size you get dozens of offers a week for paid promotions. He has to take these in order to keep his large staff media empire going. It's a vicious cycle.
I get a couple of offers week but turn them all down.

Actually, because I rarely watch LTT, is there a single video on the channel that isn't sponsored in some way?
BTW, individual video sponsorships on a channel his size would run at least 5 digits per video.

[BravoV]

The real question is, when Dave will going to debunk this at a new EEVBlog video, particularly for this LTT episode ? 

Its worth it Dave, imo.

[ogden]

AVX have "audio grade" capacitors, and they watch my videos. I'm going ask them to prove their marketing claims, and in what way their "audio grade" are manufactured differently from their regular capacitors of the exact same spec. And how they actually quantify "deeper bass" and all those wank words. Let's see if I get a response 

Meanwhile you can check AVX audio tantalum benchmark paper: https://www.interstatemarketing.com/Papers/TechArticles/Tant-Niobium/tantbench.pdf

[0culus]

As to Linus, I guess he must appeal to people who don't know any better judging on his channel size.

When your channel gets to that size you get dozens of offers a week for paid promotions. He has to take these in order to keep his large staff media empire going. It's a vicious cycle.
I get a couple of offers week but turn them all down.

Actually, because I rarely watch LTT, is there a single video on the channel that isn't sponsored in some way?
BTW, individual video sponsorships on a channel his size would run at least 5 digits per video.

Not surprising. I'm sure the guy is absolutely loaded.

[janoc]

AVX have "audio grade" capacitors, and they watch my videos. I'm going ask them to prove their marketing claims, and in what way their "audio grade" are manufactured differently from their regular capacitors of the exact same spec. And how they actually quantify "deeper bass" and all those wank words. Let's see if I get a response 

Meanwhile you can check AVX audio tantalum benchmark paper: https://www.interstatemarketing.com/Papers/TechArticles/Tant-Niobium/tantbench.pdf

You know, a benchmark "paper" (it is not peer reviewed nor published) where the authors are two reps of the capacitor manufacturer, a rep of a company selling equipment using those caps and two guys from a Czech uni that got likely roped in to give this some kind of scientific veneer (Zajaček was a PhD student and J. Šikula is likely professor Josef Šikula who was his thesis advisor - and most likely had nothing to do with this paper. It is routine to put the name of the advisor on publications by the student as a courtesy) is totally a valid and trustworthy source of information. 

And "published" by a dealer of AVX (seriously, I thought "interstatemarketing.com" was a spam domain - the name doesn't get much more spammy than this).

BTW, there is absolutely nothing about any special "audio tantalums" in that paper. They measure things like THD, sensitivity to piezoelectric effect and impact of ESR of different capacitor technologies (regular consumer tantalum, low ESR tantalum and 2 polymer tantalums in parallel, normal and low ESR niobium oxidide type, normal electrolytic and X5R MLCC) that AVX manufactures when used as a coupling capacitor for a Wolfson audio codec (supplied on an eval board - that's likely why there is the Wolfson rep as a coauthor).

I have no means to check the measurements but just looking at how this is written is making me laugh (yeah, I did review a few research papers before). The best are the final results - they recommend the tantalum caps over aluminium elcos, even though the elcos were at least as good as the tantalums in their test with this copout:

"Aluminium   capacitors   performed   well   in   the   tests,  however  a  special  care  should  be  paid  to  their  limited  reliability,  capacitance  drop  with  time  and  lead-free  process  compliance.  This  is  beyond   scope   of   this   paper. For   Reference see 7."

 

Where reference [7] is a marketing paper by AVX written by the same AVX rep that is   the coauthor of this "study" (3 out of total 8 references are self-citations by this guy) about how their niobium oxide "Oxicaps" outperform aluminium elcos. Which actually this study doesn't show - the "Oxicaps" performed the same or worse than both the tantalums and the elcos.

You can guess why they had to write that despite the results, given who has supplied the capacitors and paid for the study  I am really sorry for the PhD student that they had convinced to do the measurements and to put his name on this "paper".

(We are talking small signal audio coupling caps here, not highly loaded power supply switching filter caps or coupling caps passing many watts of power where an aluminium elco would be particularly stressed).

[ogden]

You know, a benchmark "paper" (it is not peer reviewed nor published) where the authors are two reps of the capacitor manufacturer

LOL. Do you see any other technical paper coming out of research laboratory of ANY manufacturer to be peer reviewed same way as scientific publications do? Note that AVX have capacitor research lab in Czech Republic, obviously they collaborate with University of Brno.

[janoc]

You know, a benchmark "paper" (it is not peer reviewed nor published) where the authors are two reps of the capacitor manufacturer

LOL. Do you see any other technical paper coming out of research laboratory of ANY manufacturer to be peer reviewed same way as scientific publications do? Note that AVX have capacitor research lab in Czech Republic, obviously they collaborate with University of Brno.

I am not saying it should be peer reviewed. However, then you must take it as what it is - a marketing material at best, not as a source of any hard information.

And yes, I know they have a lab in Czech republic (I think they have a factory there too). The collaboration with the Technical University Brno is also not in dispute - those people on the paper do exist (I did check who the authors are) and prof. Šikula actually works on supercapacitors and similar stuff too. That's not a problem.

The issue is that when you look at what they present in the paper in the measurement data and what they conclude at the end it doesn't make any sense. They recommend expensive capacitors over cheap elcos, despite the elcos actually performing at least as well or better than the expensive stuff (and their disadvantages being very unlikely a factor in this type of application - a typical elco will most likely outlast anything they put the capacitor in). And to suport their argument they cite materials  written by the same AVX guy.  That's totally trustworthy and reliable info, especially when their own measurements don't support the argument they are making ...

[ogden]

The issue is that when you look at what they present in the paper in the measurement data and what they conclude at the end it doesn't make any sense.

One is clear - ceramic caps are worst for both audio input and output Note that they used lowest grade MLCC (X5R). Yes, that paper is weak and not only conclusions but even test results shall be taken with grain of salt.

[janoc]

One is clear - ceramic caps are worst for both audio input and output Note that they used lowest grade MLCC (X5R). Yes, that paper is weak and not only conclusions but even test results shall be taken with grain of salt.

Oh they have actually used 2 of those MLCCs in parallel for good measure (so double the noise, microphonics and distortion, just to be sure). And even then the MLCCs performed almost identically to the rest, with the exception of the distortion below 100Hz and in the presence of vibrations (I guess they want to account for the thumping bass in the headphones ...) In other words, how to rig the setup so that the chances of getting the result you want are maximized.

[wraper]

The issue is that when you look at what they present in the paper in the measurement data and what they conclude at the end it doesn't make any sense.

One is clear - ceramic caps are worst for both audio input and output Note that they used lowest grade MLCC (X5R). Yes, that paper is weak and not only conclusions but even test results shall be taken with grain of salt.

X5R is not lowest grade, it's about the same as X7R, just with lower maximum working temperature. The worst are Z5U, Y5V.

[floobydust]

What a bunch of wankers. "Capacitor distortion" is real and most of the drama is where/how the capacitor is used in a circuit.
It's highly audible as the NFB divider (to GND) electrolytic capacitor (assuming DC bias is not a problem), and in some coupling capacitor situations.
Here are two measurements and snippets from Douglas Self and Bob Cordell's fine books which mention the topic. Elektor also did a shootout many years ago, I think it was John Lindsay Hood that wrote the article.

[TERRA Operative]

Have you looked at their price list, I couldn't find the resistors.

Edit The 0.5W Tantalum resitors are £3.39 each +Vat and the 2W up to £35.70 + Vat each.

Oof, that's expensive..... I can get Tant resistors off the shelf in Akihabara for less than a third of that IIRC.

Maybe I should get a bunch and do a heads up comparison on a class A/B amp along with carbon comp and metal film resistors using my HP 8903B audio analyzer.

[magic]

What a bunch of wankers. "Capacitor distortion" is real and most of the drama is where/how the capacitor is used in a circuit.

Yes, and "no capacitors in the signal path" is a common slogan in marketing materials of modern audio gear for that reason.
The real wankers are those who use "audio" series capacitors for power supplies, which is the remaining widespread application of electrolytics.

Wanker soundcard:
https://www.asus.com/us/Sound-Cards/Xonar_Essence_STX/
Actually, this one uses coupling capacitors despite being marketed as high end, but it's only two of them. The rest is wank.

[legacy]

Most TV shows are trash, and people do like it. Most of Youtube is completely useless trash, and people do like it.

 I am not surprised that there are guys who only care about Youtube and "Faceflop" (Facebook) rather than forums.

Youtube sounds like "easy money, honey-honey" nowadays, and there are too many idiots around who know that most users love trash.

In fact, there are even on Youtube a couple of dudes claiming they are the most brilliant minds in the universe, so brilliant and super smart that they can insult every single forum on the internet just like (they must have decided) a modern man can insult a caveman simply because the primitive communication used on forums cannot have a video stream where you watch someone.

Oh, and here if you make a video 4K, you get more credits and points.

And in this, the sad point is that with the modern Youtube you can watch bullshit, but since Youtube is not designed to facilitate "discussions", well every single bullshit remains uncommented, and people only briefly comment on his "cool" T-shirt, and "cool" desktop setup.

Does such persons deserve any better attention? Well, most people like trash, hence this is fine.

---

LTT is somehow a completely different story; they have a team, they have a forum, they discuss things ... but certain times they talk about things they don't actually know.

Yet again, because earning money on Youtube for a living require you to upload a lot of videos, sometimes you do not have the time, neither to know what you are goind to show to the audience.

[SiliconWizard]

Yep. LTT is actually not nearly as bad as many other Youtube tech channels, but the same still holds. Professional "youtubers" just sell audience. This is a form of business purely based on advertising. Making "free" content to get advertising money is nothing new; it became popular with TV, and then the Internet (I'm sure it existed before TV even, but in much less effective ways...)

The content itself just needs to be effective at keeping/increasing the audience, just like TV. Nothing else. And the content can be complete bullshit, as long as it doesn't make the audience plummet. This is the only key point.

I wouldn't say Youtube is "easy" money. It takes a lot of work to get any significant revenue from a Youtube channel. But most of this work is not  technical; which is something that eludes many engineers.

And finally, if you're looking for dependable sources of information, don't pretend that Youtube is. There's a lot of everything on Youtube. Getting to know anything about capacitors on a random Youtube channel would be like asking random people in the street. Just because someone is popular doesn't mean that they should be blindly trusted.

[legacy]

I wouldn't say Youtube is "easy" money. It takes a lot of work to get any significant revenue from a Youtube channel. But most of this work is not  technical; which is something that eludes many engineers.

I was talking about how it's perceived by most new YouTubers, especially the kind who spit against forums.

Why? Not because they believe making a video is easier than writing stuff (specifically digital reviews), but rather because they believe it's easier to persuade people with a video than with a text, this because they hate to type, to read, because they are simply ignorant, and, worse still, because they rely on ignorance to earn money.

There is a psychological mechanism behind. People tend to trust a human face rather than a PC screen with just text, therefore if you put a human face on a PC screen, you are done.

In fact, they sell "stories", which is the way these things are now "called" on twitter and similars, and stories are more about personal lifestyle and this stuff rather than on competence and technical skills, but this is what companies use nowadays to promote and sell stuff.

[SiliconWizard]

I wouldn't say Youtube is "easy" money. It takes a lot of work to get any significant revenue from a Youtube channel. But most of this work is not  technical; which is something that eludes many engineers.

I was talking about how it's perceived by most new YouTubers, especially the kind who spit against forums.

I see what you mean, but those new Youtubers are still not going to make any money (well, most of them anyway.) They can spit all they want. This won't be growing anything (except bacteria).

Oh, and seeing videos as an alternative to forums would be completely silly anyway. Videos are one-way. Forums are for discussions. There's no discussion (except with oneself) in a Youtube video. Two completely diffferent things. And don't talk about the comment section of videos. This is nothing like a discussion in most cases. "Great stuff!" and "You suck bitch!" (and variants thereof) are probably the most common comments under Youtube videos.

Of course, when there's no discussion, it makes things easier for people too lazy, and/or not clever or knowledgeable enough to discuss.

There is a psychological mechanism behind. People tend to trust a human face rather than a PC screen with just text, therefore if you put a human face on a PC screen, you are done.

Yeah. Generally speaking, we humans tend to "trust" only people we know (which kinda makes sense). And our brain shortcuts make us think that actually seeing someone makes them known to us, whereas if we have never seen them, we can't possibly know them. Sure with modern communication tools, in some cases we may actually know someone online much better than someone we've met IRL (because online they are more likely to openly share their opinions for instance), but that's all pretty recent in human history. Our brain still pretty much functions as if we were still hunters in the woods.

You'll notice that on forums. It often takes a while, and quite a few posts, before you feel like you "know" some user, and even so. If you've never seen them, you'll never really think you know them. OTOH, if you've just met someone IRL and have exchanged a few mundane sentences, you're a lot more likely to think you know them. That's just how our brain works.