Just some more peoples opinions
Measured Differences Between Capacitors for Audio Applications
Regards
M. Gregg
Measured Differences Between Capacitors for Audio Applications
Regards
M. Gregg
Measured data. Yay. Thanks for posting.
I also recall that Douglas Self measured capacitor distortion as part of his Distortion in Power Amplifiers EW/WW article series (later rewritten into a book). Cordell writes a bit about capacitor distortion as well, but leaves it at "electrolytics have higher distortion than non-polarized types". He cites the voltage coefficient of the electrolytic caps (and many ceramic caps) as the source of the distortion but doesn't elaborate.
~Tom
Joe Audiophool Will perform the test and then claim he can hear the "better" cap 50% of the time and claim that "proves" there is a difference.
BTW the only way to do this test is the have a buddy flip a coin and then either flip the switch or not. That way the listener can't know if he is hearing the same cap twice or even five times in a row. He just numbers the sessions and writes his impressions. later you look at his impressions and see if there is any corelation to which cap was being used.
I'd bet a bunch that no one who claims they can hear a cap break in actually did a test like this using a randomly selected switch setting.
Hello,
A tightly wound roll of plastic has mechanical properties, waving your hand in dismissal will not prevent mechanical resonance. It is a matter of degree, the question is, is it measurable, is it audible, is the 20 dollar capacitor better? Perhaps not! That is why we test!
Some capacitors are reported as being microphonic. Then there are condenser microphones.
DT
All just for fun!
I'm not arguing against that. I just highly doubt the resonance falls in the audio range. And I believe that's all I wrote...
I can't think of any physical gizmo that doesn't have some sort of resonance mode or another.
~Tom
Well, you make my case for me- I always recommend machine-made, mass-produced caps with very tight windings. If hard dielectrics like polypropylene are used (which allow tighter winding and are much more difficult to compress/deform), the mechanical resonances are unbelievably small.
Try this: use a mass-produced polyprop cap like a Wima in a biased RC series circuit at the input of a headphone amp. Tap the cap with a pencil. Hear anything? Now repeat this with a boutique "audiophile" Teflon cap hand-rolled by Dominican nuns. Hear anything?
Try this: use a mass-produced polyprop cap like a Wima in a biased RC series circuit at the input of a headphone amp. Tap the cap with a pencil. Hear anything? Now repeat this with a boutique "audiophile" Teflon cap hand-rolled by Dominican nuns. Hear anything?
Some older ceramic dielectric capacitors are well known to be microphonic, along with poor connections and faulty resistors.
It was good to see that Conrad Hoffman was able to set up a test system for showing relative residual capacitor distortion and presenting results at 1kHz and 3.6Vrms applied amplitude for a good range of 10nF caps.
Of interest to me is getting a better definition of the distortion from the relative high residual capacitors, as his present system is not robust enough for discerning low level residuals. It would be great if he could one day extend his measurements to other frequencies and amplitudes, especially for the capacitor samples with the highest level of residual, and do a harmonic assessment. He may even be able to extend his measurement to capacitances in the uF range and include a DC bias for electrolytic capacitor residual assessment. If so, then that would be an appropriate tool to investigate whether 'burn-in' had an effect on the distortion generating mechanism of the cap.
One possible influence for noticing a change in amplifier sound character over time is part temperature, where some parts can take many hours to reach thermal stability. There may also be other electrolytic cap related inlfuences that don't show up in a bridge residual test system, such as varying capacitor leakage current as a function signal voltage, or dielectric absorption, and its influence on the surrounding circuit.
Designing Audio Power Amplifiers By Bob Cordell presents and discusses capacitor distortion test results.
Menno van der Veen & Hans van Maanen have a 2008 AES paper discussing a certain capacitor distortion mechanism, but I haven't seen the paper.
Renardson has some discussion at:
CAPACITORS
Curl and Jung prepared the paper in
http://waltjung.org/PDFs/A_RealTime_Signal_Test_For_Capacitor_Quality.pdf
It was good to see that Conrad Hoffman was able to set up a test system for showing relative residual capacitor distortion and presenting results at 1kHz and 3.6Vrms applied amplitude for a good range of 10nF caps.
Of interest to me is getting a better definition of the distortion from the relative high residual capacitors, as his present system is not robust enough for discerning low level residuals. It would be great if he could one day extend his measurements to other frequencies and amplitudes, especially for the capacitor samples with the highest level of residual, and do a harmonic assessment. He may even be able to extend his measurement to capacitances in the uF range and include a DC bias for electrolytic capacitor residual assessment. If so, then that would be an appropriate tool to investigate whether 'burn-in' had an effect on the distortion generating mechanism of the cap.
One possible influence for noticing a change in amplifier sound character over time is part temperature, where some parts can take many hours to reach thermal stability. There may also be other electrolytic cap related inlfuences that don't show up in a bridge residual test system, such as varying capacitor leakage current as a function signal voltage, or dielectric absorption, and its influence on the surrounding circuit.
Designing Audio Power Amplifiers By Bob Cordell presents and discusses capacitor distortion test results.
Menno van der Veen & Hans van Maanen have a 2008 AES paper discussing a certain capacitor distortion mechanism, but I haven't seen the paper.
Renardson has some discussion at:
CAPACITORS
Curl and Jung prepared the paper in
http://waltjung.org/PDFs/A_RealTime_Signal_Test_For_Capacitor_Quality.pdf
I have. They consider something analogous to electrostriction with compression of the dielectric from the applied field. There's some fundamental flaws, most notably the lack of scaling or order of magnitude calculations for the actual physical compression of polymer films in caps due to their proposed mechanism.
Jantzen audio claims that they developed very stiff winding for their caps, so they dont neeed to burn in.
So. Manufacturer of caps claims that the winding method is important factor!
I have bunch of Jantzen Superiors infested all around my hi fi system and they really need maybe a day to show out.
Russian PIOs, Mundorf SIOs, i have also, they are a different story. They need much more to burn in.
Different chemical components in different types of caps, more, or less stiff winding, those are two proven factors to answer the question, without need to fill the sophas..
For non polar caps i use slow method for dummies.
I just add a bunch of them on a speaker driver in "silent mode"- equipped with a large resistor to totally silent it and after some time i "harvest them" for normal use.
So. Manufacturer of caps claims that the winding method is important factor!
I have bunch of Jantzen Superiors infested all around my hi fi system and they really need maybe a day to show out.
Russian PIOs, Mundorf SIOs, i have also, they are a different story. They need much more to burn in.
Different chemical components in different types of caps, more, or less stiff winding, those are two proven factors to answer the question, without need to fill the sophas..
For non polar caps i use slow method for dummies.
I just add a bunch of them on a speaker driver in "silent mode"- equipped with a large resistor to totally silent it and after some time i "harvest them" for normal use.
Last edited:
Almost all ceramic capacitors exhibit some amount of piezoelectric effects. Some are quite bad. It may not be discussed in the data sheet. The effect goes both ways.
If the capacitor is used in a sensitive circuit, say the loop filter of a phase locked loop, the capacitor itself will be able to modulate the VCO. If said capacitor was SMD the entire PC board becomes a microphone and the PLL becomes an FM transmitter.
If the capacitor is used as a typical B+ filter, in a circuit with rapidly changing currents, say the RF PA B+ bypass capacitor in a TDMA cell phone (The transmitter pulses on and off in TDMA), then the entire PC board will buzz or tick at the TDMA rate.
Both are actual occurrences from my past. Break in will not help. The piezoelectric effect is a property of the ceramic material used.
Special non - piezo capacitors are available but the dielectric constant of the material is lower so the capacitor is much bigger for a given value. Often a tantalum capacitor is used instead.
Yes we often refer to a tube, or a capacitor, as being 'microphonic', if the experience is blatently obvious. Even NPO would be microphonic if the application could portray it sufficiently well. Part fabrication, and application techniques, change for the better and the affect typically becomes less noticable as the years roll on. And then a new application, or new generation rolls along, and the topic comes up again!
I believe he is saying that people can't tell the difference between very good amplifiers, as they all distort the signal by such a small amount that it is not noticeable. He is also saying that many people do not like the sound of these amplifiers as they prefer some distortion, so they choose a capacitor which gives them the type/amount of distortion they prefer. (Having done this, they then proclaim that their favourite distortion is not distortion).
In general, you can't distinguish between different caps in a very good amp as the amp performance will not be set by the caps but by the feedback resistors. I think he may be saying that the sort of people who worry too much about caps (beyond normal engineering issues such as known non-linearities) would not wish to listen to a very good amp anyway.
In general, you can't distinguish between different caps in a very good amp as the amp performance will not be set by the caps but by the feedback resistors. I think he may be saying that the sort of people who worry too much about caps (beyond normal engineering issues such as known non-linearities) would not wish to listen to a very good amp anyway.
I wouldn't put much faith in a manufacturer's claims. I have seen product bugs or oddities turned into features by clever marketing and advertising. Also, if the Audiophool market is big enough and enough people worry about break-in of their capacitors, it would be a value-adder to sell capacitors that were already broken in or didn't require break-in at all. My reaction is "yeah, whatever"...
I'm also highly skeptical of speaker driver manufacturer's claims that their drivers should be broken in with certain kinds of music during the first X hours and then another kind of music for Y hours. Speaker break-in is a matter of getting the suspension loosened up. If it's done with heavy metal, Beethoven, or a sine wave shouldn't matter.
~Tom
I connected my recently built speakers up to an old Panasonic boom box and cranked it up for 10 hours a day while I went to work. I pushed the bass slider up way into the distortion zone too. I used the boom box because I didn't want to leave a tube amp on while not home. I could hear small differences in the sound after each day and the speakers seem to get louder too. I noticed the biggest change when I left the boom box on a rap music station, but I only did this once because I didn't want to polute the speakers. Most days they got classic rock. I think the rap crap just produced more cone movement, so it acted faster, but in the end anything would have worked.
I picked up a pair of TangBand full-range drivers on sale recently. I ran them close to their maximum excursion using a 10 Hz sine wave for the better part of 24 hours. I haven't actually bothered building boxes (or in my case spheres) for them yet... Wouldn't want to rush into things ya' know. I have a tube amp to finish first.
I used 10 Hz so I didn't have to listen to the drivers while they broke in.
~Tom
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.