I recently received an article from Wurth Electronics on the acoustic effect of Harmonic Distortion caused by Aluminum Electrolytic capacitors. Their findings pretty much agree with what I have observed and measured over the years - decades actually. This information isn't unique to Wurth products, it applies to all electrolytic capacitors.
I contacted representatives at Wurth for permission to post the article here. They agreed after contacting the author, Dr. Rene Kalbitz.
... so here it is as published.
I contacted representatives at Wurth for permission to post the article here. They agreed after contacting the author, Dr. Rene Kalbitz.
... so here it is as published.
For those who do not wish to wade through the 15 page pdf:
"Abstract: This note reports a comparative study of total
harmonic distortions (THD) caused by commercial electrolytic
capacitors, as produced by Würth Elektronik eiSos as well as
purpose-built items. The discussion about the audibility of
distortions is made on the basis of human sound perception.
This note arrives at the conclusion that capacitors do not add
significant distortions to fundamental frequencies as they
transfer signals. Modifications of the electrolyte or
separation paper have almost no effect on the THD."
"Abstract: This note reports a comparative study of total
harmonic distortions (THD) caused by commercial electrolytic
capacitors, as produced by Würth Elektronik eiSos as well as
purpose-built items. The discussion about the audibility of
distortions is made on the basis of human sound perception.
This note arrives at the conclusion that capacitors do not add
significant distortions to fundamental frequencies as they
transfer signals. Modifications of the electrolyte or
separation paper have almost no effect on the THD."
Within reason, and assuming reasonable materials and construction, yes. That is correct. The products made by Wurth are very much like other brands (even special ones for audio), so this applies to electrolytic capacitors in general. If you read the note, they did vary materials and note the results.
There is more strange things. Is threshold of audible distortions really 10 % on 55.8 Hz and 448.9 Hz?
I added the harmonics:
Saw: 62.4%
Square: 43.2%
Triangle: 12.0%
Sine: 0%
The study is considering only non-linear capacitance and ignoring how the capacitor is used in a circuit. A 500 ohm series resistor will dilute the 470uF capacitor's THD to essentially zero.
Ed
Saw: 62.4%
Square: 43.2%
Triangle: 12.0%
Sine: 0%
The study is considering only non-linear capacitance and ignoring how the capacitor is used in a circuit. A 500 ohm series resistor will dilute the 470uF capacitor's THD to essentially zero.
Ed
Didn't Doug Self say something about this some years ago? IIRC he found that if the AC voltage across an electrolytic was less that 60 some odd mV then there was no measurable HD?
IIRC the conclusion therefore reached by some here in the forum was that if the cap was a large enough value such that at the lowest frequency of interest the voltage across the cap below 60mV or so then the cap should have no effect on distortion.
IMHO it simply raises the question as to what makes different kinds of electrolytic caps in sensitive parts of some circuity sound the way they do. Okay, so not HD. Then its just one or more other already known physical phenomena that must be responsible.
IIRC the conclusion therefore reached by some here in the forum was that if the cap was a large enough value such that at the lowest frequency of interest the voltage across the cap below 60mV or so then the cap should have no effect on distortion.
IMHO it simply raises the question as to what makes different kinds of electrolytic caps in sensitive parts of some circuity sound the way they do. Okay, so not HD. Then its just one or more other already known physical phenomena that must be responsible.
I think this is common knowledge. The capacitors that affect the sound have reactance at audio frequencies - equalization networks and crossovers.
Ed
Ed
That is exactly the entire point. A coupling capacitor does not have signal voltage developed across it. Otherwise it becomes partly a filter at least. If you size the capacitor properly, it has to be pretty bad quality to make an audible difference. Read defective.
The work Doug Self did was on track and excellent. Other similar studies came to the same conclusions.
As for the sound of various capacitors. Well, that has a great deal more to do with expectation bias and other tricks of the mind. I've tested this often enough over the years using the average customer as a test bed. People who hear a difference aren't really hearing things, but their mind has been conditioned to sense a difference. So they do.
The work Doug Self did was on track and excellent. Other similar studies came to the same conclusions.
As for the sound of various capacitors. Well, that has a great deal more to do with expectation bias and other tricks of the mind. I've tested this often enough over the years using the average customer as a test bed. People who hear a difference aren't really hearing things, but their mind has been conditioned to sense a difference. So they do.
Hi Mark,
Trying to put real numbers or limits on this is problematic. I don't know if you can set a level where everyone will or will not hear something is possible.
I really do not want things to get bogged down in that kind of argument, that isn't the point of the discussion.
Trying to put real numbers or limits on this is problematic. I don't know if you can set a level where everyone will or will not hear something is possible.
I really do not want things to get bogged down in that kind of argument, that isn't the point of the discussion.
Haha. Not talking about FR, HD, nor differences in reactance. Its the other things. Muddyness, false edgy-brightness, etc. These aren't necessarily HD, nor FR effects.
For instance, an electrolytic coupling cap can be made arbitrarily large, but one brand/model still sounds different from another. ESR could be one factor, but it doesn't seem to be the only factor. For electrolytics, could be current noise is an issue. But nobody measures electrolytic cap current noise, so its its outside of WYSIATI.
Anyway, if we want to limit the discussion to measured HD only, then I would agree its a trivial thread.
Last edited:
That factor is reliability. High-end equipment avoids electrolytic capacitors because they are unreliable.
Ed
Ed
Mark, this is then a trivial thread for you. It's about the physics and science behind transmitting a signal.
Your entire premise is pure speculation, undocumented by any type of measurement or real controlled testing. Controlled testing through listening has disproved your ideas here. So let's not devolve this thread into arguments more related to religion than anything that can be proved by any accepted proper testing methods.
You know, if what you are claiming had been proved to exist by any accepted testing, I'd be fully on board. But designing a test to exclude outside factors is extremely difficult, it requires skill and careful setup and running of the test. It has been done, but some folks do not like the results. I suspect you are in that group. I myself have listened carefully over the decades, I have yet to hear a properly sized coupling cap in good condition be audible in any way.
Your entire premise is pure speculation, undocumented by any type of measurement or real controlled testing. Controlled testing through listening has disproved your ideas here. So let's not devolve this thread into arguments more related to religion than anything that can be proved by any accepted proper testing methods.
You know, if what you are claiming had been proved to exist by any accepted testing, I'd be fully on board. But designing a test to exclude outside factors is extremely difficult, it requires skill and careful setup and running of the test. It has been done, but some folks do not like the results. I suspect you are in that group. I myself have listened carefully over the decades, I have yet to hear a properly sized coupling cap in good condition be audible in any way.
Chris,
I could post schematics of high-end equipment that uses electrolytic caps in the signal path. Its nonsense to claim they are not used in high end equipment. You know that. Sometimes they are the least-worst solution.
Also, most controlled listening tests are amateur bunk. They are very rarely performed according to a modern understanding of perceptual science.
However its your thread to do with as you wish.
Good day to you sir. I am out.
I could post schematics of high-end equipment that uses electrolytic caps in the signal path. Its nonsense to claim they are not used in high end equipment. You know that. Sometimes they are the least-worst solution.
Also, most controlled listening tests are amateur bunk. They are very rarely performed according to a modern understanding of perceptual science.
However its your thread to do with as you wish.
Good day to you sir. I am out.
Last edited:
Current noise is measured by the way. It is very important in electronic testing. It is dependent on the circuit and applied voltages. So many factors affect this, temperature being one huge factor. You test in the application, but leakage amounts (on the data sheet) give clues.
Hi EdGr,
Electrolytic capacitors are pretty reliable these days. They have a "best before date" in run time hours, but so do many other parts. I wouldn't go so far as to say they are unreliable, but maybe they are not as reliable as some other parts.
One thing that is true. In so called "high end" equipment that runs hot, smaller capacitors will tend to fail earlier. That's because failure rate doubles for every 10 °C rise in ambient temperature, and there is less electrolytic in smaller caps, and the seals aren't as good. Now, knowing this, doesn't it follow that equipment that runs hot is by design, unreliable compared to equipment running at more normal temperatures?
Just something to consider.
Hi EdGr,
Electrolytic capacitors are pretty reliable these days. They have a "best before date" in run time hours, but so do many other parts. I wouldn't go so far as to say they are unreliable, but maybe they are not as reliable as some other parts.
One thing that is true. In so called "high end" equipment that runs hot, smaller capacitors will tend to fail earlier. That's because failure rate doubles for every 10 °C rise in ambient temperature, and there is less electrolytic in smaller caps, and the seals aren't as good. Now, knowing this, doesn't it follow that equipment that runs hot is by design, unreliable compared to equipment running at more normal temperatures?
Just something to consider.
I suppose reliability is relative. The capacitors I have seen fail were large filter capacitors after 30 years. That could be considered reliable for many products.
Ed
Ed
Hi Ed,
In practice, I see some large filter capacitors fail. However far more smaller capacitors fail. I will agree that electrolytic capacitors are one of the least reliable components used, Tantalum capacitors being even less reliable (often with spectacular results!).
Of the large capacitors I see fail, most were 40 + years old, or they were inexpensive filter capacitors. High ambient temperatures really accelerate failure, filter caps in a Krell being an easy example of this. Another major factor is not enough headroom between actual applied voltage and the rating.
None of this is really surprising.
Consider too, large value filter capacitors are expensive. So what you might see in equipment designed for consumer use is the least expensive they can get away with. That includes switching power supply products like TVs. I have a Marantz 500 sitting here with good filter capacitors. They stopped production in 1974.
In practice, I see some large filter capacitors fail. However far more smaller capacitors fail. I will agree that electrolytic capacitors are one of the least reliable components used, Tantalum capacitors being even less reliable (often with spectacular results!).
Of the large capacitors I see fail, most were 40 + years old, or they were inexpensive filter capacitors. High ambient temperatures really accelerate failure, filter caps in a Krell being an easy example of this. Another major factor is not enough headroom between actual applied voltage and the rating.
None of this is really surprising.
Consider too, large value filter capacitors are expensive. So what you might see in equipment designed for consumer use is the least expensive they can get away with. That includes switching power supply products like TVs. I have a Marantz 500 sitting here with good filter capacitors. They stopped production in 1974.
Hi Mark,
I work on equipment every day that use many electrolytic caps. I agree with you on this, and I think you were actually talking to someone else.
Agreed. Amateur testing is typically not even close to controlled. I am talking about real, carefully designed and controlled tests done many times over the years. Exactly the same as we do with anything else in life.
"My thread"? I guess I started it. I just do not want it to get dragged off into the weeds because the message gets lost when that happens. If there is actual properly tested and documented information out there, anyone is free to post it. Proper testing means it is reproducible by others.
I work on equipment every day that use many electrolytic caps. I agree with you on this, and I think you were actually talking to someone else.
Agreed. Amateur testing is typically not even close to controlled. I am talking about real, carefully designed and controlled tests done many times over the years. Exactly the same as we do with anything else in life.
"My thread"? I guess I started it. I just do not want it to get dragged off into the weeds because the message gets lost when that happens. If there is actual properly tested and documented information out there, anyone is free to post it. Proper testing means it is reproducible by others.