FYI - here's a link to an article often referred to in the DIY community showing the measurement of capacitor linearity. Anyone with some basic test equipment can duplicate this.
The "Sound" of Capacitors
But, again, note a rather large "AC" voltage drop is employed to see the effect. Those so inclined could also do some general research on capacitor paraelectric properties.
It's probably best I stop here to avoid any more "wrath". Just wanted to provide more info for those who want to dig a bit deeper into this somewhat controversial subject. I admit to being more in the "objectivist" camp.
FYI - the work of Mr. Bench was cited several times along the first thread.
As the semi-axes of the ellipse are
a ≈ (V/d) ω
b ≈ (V/d) ∣ε(ω)∣ ESR ω C
If you want to see something on the scope, you must increase the applied AC voltage for a given frequency.
As the laws of physics can't be broken, I consider that my equations are quite "objectivists"
Last edited:
Yes, but my point was there should be almost NO AC "signal" voltage dropped across an audio "coupling" capacitor if it is properly sized to pass the frequency band of interest. Applying unrealistically large AC voltages across a capacitor just to reveal it has some non linearity is an interesting exercise, but not an appropriate analysis for audio coupling applications. I only included the link to Steve's page to show that non linearity can be easily demonstrated. Note, though, that capacitors used in speaker crossover networks are NOT coupling capacitors, rather filter capacitors where it is expected that voltage will be dropped. Different subject.
I fully believe, and appreciate, that you are taking the "objectivist" point of view and backing it with physics. Perhaps it could be we just have a disconnect in communication? Would you agree that "if a passive component passes an audio signal and drops no AC signal voltage in the process (thus acts as a short) it therefore cannot introduce any distortion? Also, if a component does drop some voltage, but is perfectly linear in operation, then it will only decrease the level of the signal, but not introduce any distortion. Finally, if a component does drop some signal voltage and exhibits non linearity, then it not only reduces the signal level, it will also introduce some distortion".
And thus, a coupling capacitor, undersized for the application, will drop some "AC" signal voltage and, if it exhibits any non linearity, may also introduce some distortion. So it follows that, "properly sized" coupling capacitors, of any type, or brand, should theoretically not introduce any significant signal reduction, or distortion, or exhibit audible differences. That's basically my premise.
Interesting,
So phase shift is not frequency dependant.. a coupling capacitor does not just pass a single frequency so if several frequencies (a complex signal) passes through a capacitor there will be no effect..it will be just as perfect as it as it was when created. This would assume that the phase shift would be the same across the frequency range.
Of course if there was a frequency dependant shift that would be what is known as smear...(but only by the audiophools).
Comments like it sounds like I have taken my speaker from behind a curtain and it sounds so much clearer now just like I have removed a rag from the speaker cone.. 😀..
No I think its all imagination to be honest..how can a capacitor remove a rag from a speaker cone..
And this capacitor is great at vocal music but not good with pop..anyone would think the smear was frequency dependant..its not very good with complex music..but the vocals shine through..😀
Regards
M. Gregg
So phase shift is not frequency dependant.. a coupling capacitor does not just pass a single frequency so if several frequencies (a complex signal) passes through a capacitor there will be no effect..it will be just as perfect as it as it was when created. This would assume that the phase shift would be the same across the frequency range.
Of course if there was a frequency dependant shift that would be what is known as smear...(but only by the audiophools).
Comments like it sounds like I have taken my speaker from behind a curtain and it sounds so much clearer now just like I have removed a rag from the speaker cone.. 😀..
No I think its all imagination to be honest..how can a capacitor remove a rag from a speaker cone..
And this capacitor is great at vocal music but not good with pop..anyone would think the smear was frequency dependant..its not very good with complex music..but the vocals shine through..😀
Regards
M. Gregg
Last edited:
😱..but what would happen if its connected to different circuits..
The cap performance would be circuit dependant...😀
No its all rubbish...circuit dependant..performance related to everything else..all imagination you will be using coupling transformers next..but they aren't very good at..yawn.😴
Regards
M. Gregg
The cap performance would be circuit dependant...😀
No its all rubbish...circuit dependant..performance related to everything else..all imagination you will be using coupling transformers next..but they aren't very good at..yawn.😴
Regards
M. Gregg
Last edited:
If the cap is properly sized for its role as a coupling cap (that means much larger than what is needed to have a highpass at 20Hz) then LF phase shift is negligible (and doubly so compared to LF phase shift od bass reflex speakers)...
Yes, this is primarily what matters for "coupling" capacitors. However, there are also other interesting things people often don't consider. Specialty "boutique" capacitors, of a given value, are often much larger "physically" compared to the OEM parts they are intended to replace. This often results in the new capacitors hanging in the air, on long leads, since they no longer fit on the circuit board, or in the space, the originals occupied. This may result in noise pickup, or instability, that never existed before.
The question of "phase shift" as per the above comments, is not an issue in the frequency band of interest. Of course, phase shift will occur as the frequency transitions from 0Hz to the lowest the lowest frequency of interest (usually 20Hz) since it acts a a high pass filter in that region. But beyond that, phase shift ranges from negligible to zero, as the frequency increases. While some will try to make an issue of various other internal parasitic effects, these are insignificant at audio frequencies. The audibility of phase shift is yet another controversial subject. I find it interesting, however, that some people will go to great lengths to eliminate any such possibility in their amplifiers (bypassing tone controls, for example) and yet, are perfectly content with the large phase shifts that occur as a result of their speaker crossover networks.
well looking at this,
Testing audiophile claims and myths
I think we should all pack up and go home, in blind tests you can't tell any amp from any amp any system "reliably" from another and if its Valve /transistor/ chip/ or a combination of the two..
So there you go..buy a Midfi system and or a DAB radio and your in audio nirvana.
The question could also be asked if you were blindfolded could you tell what kind of car you were in on a short trip down a road..
9 out of 10 cats prefer steak but who knows whats in the cat food! (Steak on the tin must be steak)
If this is really true then measurements mean nothing listening tests mean nothing and circuit types don't mean a great deal either..
Regards
M. Gregg
Testing audiophile claims and myths
I think we should all pack up and go home, in blind tests you can't tell any amp from any amp any system "reliably" from another and if its Valve /transistor/ chip/ or a combination of the two..
So there you go..buy a Midfi system and or a DAB radio and your in audio nirvana.
The question could also be asked if you were blindfolded could you tell what kind of car you were in on a short trip down a road..
9 out of 10 cats prefer steak but who knows whats in the cat food! (Steak on the tin must be steak)
If this is really true then measurements mean nothing listening tests mean nothing and circuit types don't mean a great deal either..
Regards
M. Gregg
Last edited:
If accepted circuit theory and psychoacoustics say that two things should sound about the same, and unsighted testing shows that they do appear to sound about the same, why are people surprised and predict the end of the world etc.?
If you take the trouble to read, derived equations work for AC voltages across the capacitor.
BTW, to have a hysteresis loop with DC voltage it is a bit trickier.
A DC offset has the only effect of moving hysteresis loop along E-D plane and avoid zero crossing if it is high enough.
Can you please explain how and where?
It is said out there that a 1 μF film/foil polypropylene capacitor gives about -130dB distortion, I wonder how easy is to measure that.
My own analysis is based on solid physical foundations, and it is quite logic.
Maybe it has not been discussed well enough?
Although not very appropriate from a moderator, but funny quote BTW.
I must assume that you are the teacher?
Last edited:
Are you assuming that "real" capacitors behave in a linear manner regardless of construction, or operating conditions? Sorry, but they don't.
I supplied the link to Steve's page which demonstrates the effect well. You should have noticed the pronounced S-curve for the ceramic capacitors tested. This is the expected result when the "AC" voltage across certain ceramic capacitors is great enough for them to exhibit ferro electric affects. The capacitor "changes value" under these conditions.
Ceramic capacitors, especially those of the X7R and Z5U type, exhibit this effect to a very large degree, and thus is easy to see under simple test conditions, as performed by Steve. Ceramic capacitor manufacturers often supply information showing such changes right on their data sheets. Here's a link to a data sheet for a Kemet 0.1uF capacitor showing just that. Scroll to page 4 and observe the graph clearly showing the how capacitance changes value significantly with AC voltage drop.
http://www.kemet.com/Lists/ProductCatalog/Attachments/58/F3101_Aximax.pdf
I invite you to work that into your calculations!
Ceramic capacitors can exhibit non linearity to a large degree, but all "real world" capacitors exhibit some non linearity. The only linear dielectric is a vacuum. Film capacitors are respectably linear, for the most part. Ceramics and polarized capacitors are a lot less so. And, yet again, when properly sized, "coupling" capacitors drop virtually no AC voltage, thus minimizing dielectric non linearity effects.
Definitely you didn't read my posts, I said almost the same that you say, except that capacitor linearity (non-linearity) has nothing to do with AC applied voltage, but with intrinsic dielectric properties and capacitor geometry.
However, some capacitors are more linear than any valve, transformer or transistor.
Let's try to not feed another audio myth. 😉
I invite you to read that into my calculations!
http://www.diyaudio.com/forums/lounge/207544-ecc82-12au7-line-preamp-36.html#post3974352
The change in capacitance with applied AC voltage in ceramic capacitors is due to electrostriction and/or piezoelectric effect, both ignored in my calculations, for simplicity.
For the rest, I totally agree with you. 🙂
Last edited:
Yes, I read your posts and was certainly confused when, on the one hand, you did seem to acknowledge capacitor dielectric non linearity, but then, apparently, did not seem to think it was measurable, or demonstrable. Indeed, I must admit, I think some confusion remains when still you say capacitor linearity is not affected by AC voltage drop, when even manufacturers data sheets show it can be. A component that "changes its value" under varying operating conditions is the very essence of non linearity. Certainly, if one were to use that Kemet capacitor I referenced as an "undersized" coupling capacitor, frequency response could be quite non linear under certain conditions. Yet, capacitors properly chosen, and sized for the application, are indeed more linear than any valve, transformer, or transistor. I think we agree!
At this point, I'm going to consider that we are generally in agreement. There may just have been a bit of a communications gap, perhaps on my part. And, I really, REALLY, do not want to perpetuate any audio myths. I don't buy into any of them. Especially "magic" audio coupling capacitors at outrageous prices.
I think this little trek has run its course and thank you for your "objective" input. 🙂
At this point, I'm going to consider that we are generally in agreement. There may just have been a bit of a communications gap, perhaps on my part. And, I really, REALLY, do not want to perpetuate any audio myths. I don't buy into any of them. Especially "magic" audio coupling capacitors at outrageous prices.
I think this little trek has run its course and thank you for your "objective" input. 🙂
I must to apologize, my equations arise in rare moments of inspiration, even worse, I think slowly, then sometimes I left some things as understood. 😱
Ceramic capacitors exhibits mostly pathological behavior, and ferroelectric effect, piezoelectric effect and electrostriction make me left them aside into my calculations, just for simplicity.
The work of Mr. Bench is one of the most serious around there and it was a kind of inspiration, Lissajous figures AKA Bowditch curves, brilliant!
First time that I derived the equations, I could not believe, so I checked them several times, and they are quite consistent with results of Mr. Bench, at least for well behaved dielectrics.
I myself use cheap industrial grade MKP capacitors with amazing results, and differences with boutique capacitors are mostly marginal, sometimes favoring cheap caps due to geometry, inductance, etc. (By datasheet comparison)
Then, objectively, we can say that we agree in almost all. 🙂
The tire manufacturing companies do this all the time!
At the obstacle course/test track, they put the test tires on a nice sedan, cover the tires with bags, the test driver goes around the course with some timed sections and some subjective sections. When the driver comes back to the garage they cover the tire and he goes into the office to write a subjective report while they put on other tires. This goes on all day, at some points they will retest tires. Both the timed sections and the subjective report better match or that driver is out of a job.
- Status
- Not open for further replies.