So for example, with numbers that I pluck out of thin air:
You apply a DC bias voltage and a 1 kHz signal voltage to a capacitor.
The plates start moving. The movement repeats 1000 times per second, but as the waveform is not sinusoidal, it contains higher harmonics.
Due to a sharp mechanical resonance at (for example) 6 kHz, the 6 kHz component is unusually strong compared to the other harmonics.
The mechanical vibrations modulate the capacitance, so in the electrical current, you get sum and difference products of the 1 kHz electrical signal and the 6 kHz movement.
Hence, you have harmonic distortion with relatively strong 5th and 7th harmonics.
It's still harmonic distortion, but its level might have weird frequency dependencies and it might have a weird harmonic distribution, if these phenomena are strong enough to dominate the distortion. For example, if the supposed sharp resonance would be at 5.4 kHz, you wouldn't notice it with a 1 kHz signal, but you would with 900 Hz.
I don't remember if Cyril Bateman did any sweeps over frequency. If he or any of the others who measured capacitor distortion did do sweeps over frequency, they must have encountered the hypothetical mechanical-resonance-enhanced distortion, if it exists.
You apply a DC bias voltage and a 1 kHz signal voltage to a capacitor.
The plates start moving. The movement repeats 1000 times per second, but as the waveform is not sinusoidal, it contains higher harmonics.
Due to a sharp mechanical resonance at (for example) 6 kHz, the 6 kHz component is unusually strong compared to the other harmonics.
The mechanical vibrations modulate the capacitance, so in the electrical current, you get sum and difference products of the 1 kHz electrical signal and the 6 kHz movement.
Hence, you have harmonic distortion with relatively strong 5th and 7th harmonics.
It's still harmonic distortion, but its level might have weird frequency dependencies and it might have a weird harmonic distribution, if these phenomena are strong enough to dominate the distortion. For example, if the supposed sharp resonance would be at 5.4 kHz, you wouldn't notice it with a 1 kHz signal, but you would with 900 Hz.
I don't remember if Cyril Bateman did any sweeps over frequency. If he or any of the others who measured capacitor distortion did do sweeps over frequency, they must have encountered the hypothetical mechanical-resonance-enhanced distortion, if it exists.
Last edited:
Ok, here is another example. A capacitor has one of intrinsic resonances at 5 kHz. You feed it signal of 4.6 kHz. Because resonance is broad (follows Gauss curve), the 4.6 kHz input will excite 5 kHz oscillation, which is a non-harmonic distortion.
Do the test mentioned above and you will be convinced that mechanotransducer distortion in capacitors really exists. THD sweep across audio spectrum would show anomalously high distortion at certain frequencies. Subjectively, the sound is akin to that of a buzzing old time metallic grammophone membrane.
Do the test mentioned above and you will be convinced that mechanotransducer distortion in capacitors really exists. THD sweep across audio spectrum would show anomalously high distortion at certain frequencies. Subjectively, the sound is akin to that of a buzzing old time metallic grammophone membrane.
@sser2, You seem to have spent some time listing to capacitors. IME you make some good points.
BTW, John Curl told me its best to design out any coupling caps you if you can. Good ones cost too much and they still color the sound.
Electrolytic caps seem to have more variation in terms of how they affect sound, and it tends to change as they age. Sometimes there can be some HD, sometimes there can be signal-correlated noise (and not necessarily simple white noise either), sometime frequency dependent 'blooming' at LF, etc. Apparently something to do with the chemistry of specific electrolyte types which may change somewhat as they age.
Film caps are more simple, but, yes, they can be electromechanical transducers if not tightly wound. Physically big ones can act like antennas. Different dielectrics reach DA equilibrium under end use operating conditions over time to produce different sounds, some are bright, some are more subdued, etc. The most neutral caps I have found are the Rel polystyrene and foil ones originally patented by Richard Marsh (RTX). I was told that low voltage polystyrene film of the original type is apparently no longer avaiable, so small body size RTX caps of a few uf are no longer made. Only retail source I know of is Michael Percy. https://www.percyaudio.com/Catalog.pdf
At least Rel RTX are as uncolored and neutral as any caps I have tried, but they can still take some days to equilibrize under operating conditions. Sensitive midrange sounds open up a little more and sound is less congested, or one could say individual sounds are less smeared together. Its a subtle effect, some people won't notice what is changing, other people easily notice.
As @lrisbo said about an audible effect in electromagnetic speakers of the type he works on, although blind tests show typical listeners don't hear the effect, everyone on the design team can hear it. They think anyone can learn to hear it. IME the learning to hear it possibility may be true for a lot of the stuff that people argue about. My advice: stop trying to listen for what you know about. Don't listen for FR or for HD, etc. Just listen to, say, one dac verses another dac without trying to notice anything technical in particular. You might focus on little details of the music, how piano cords form textures. Maybe how cymbals sound, like bursts of noise, like rushing water, like sizzling grease, like a real cymbal struck with a stick with a plastic tip, or maybe with a wood a wood tip, etc. A whole lot of detail is there if one listens and if the reproduction system is good enough.
I might add that a lot of small audible effects are also easy to mask. Listening on an unfamiliar system can be disorienting as well. Adding DBT relays and switches can add masking distortions or not. Make sure the designer of such systems knows how to make sure the the test gear does not mask the effects being tested for.
For anyone who has practiced mindful meditation, there may be an analogy to mindful listening. One trick to achieving a mindful meditation state is to attempt to be consciously be aware of sensation of every part of one's body all at once. its done by starting with the toes working up the legs, and adding awareness of each little part while dropping awareness of none. If done properly the amount of information is overwhelming. It easily fills up conscious awareness. However it takes practice not to drop awareness of some already sensed parts as new parts are added. So it is with listening to music. The brain filters out much of existing but nonessential sensory information automatically, particularly if we are busy thinking about FR, HD, or anything thing else in particular.
Anyway, maybe that's one way to think about the problem of trying to learn to develop a mindful type of listening. It may take some practice filling the mind with attention to a lot of little details of musical sounds, and little details of the sound of a space where music is recorded. On mindful meditation: https://www.frontiersin.org/articles/10.3389/fnhum.2013.00012/full
BTW, John Curl told me its best to design out any coupling caps you if you can. Good ones cost too much and they still color the sound.
Electrolytic caps seem to have more variation in terms of how they affect sound, and it tends to change as they age. Sometimes there can be some HD, sometimes there can be signal-correlated noise (and not necessarily simple white noise either), sometime frequency dependent 'blooming' at LF, etc. Apparently something to do with the chemistry of specific electrolyte types which may change somewhat as they age.
Film caps are more simple, but, yes, they can be electromechanical transducers if not tightly wound. Physically big ones can act like antennas. Different dielectrics reach DA equilibrium under end use operating conditions over time to produce different sounds, some are bright, some are more subdued, etc. The most neutral caps I have found are the Rel polystyrene and foil ones originally patented by Richard Marsh (RTX). I was told that low voltage polystyrene film of the original type is apparently no longer avaiable, so small body size RTX caps of a few uf are no longer made. Only retail source I know of is Michael Percy. https://www.percyaudio.com/Catalog.pdf
At least Rel RTX are as uncolored and neutral as any caps I have tried, but they can still take some days to equilibrize under operating conditions. Sensitive midrange sounds open up a little more and sound is less congested, or one could say individual sounds are less smeared together. Its a subtle effect, some people won't notice what is changing, other people easily notice.
As @lrisbo said about an audible effect in electromagnetic speakers of the type he works on, although blind tests show typical listeners don't hear the effect, everyone on the design team can hear it. They think anyone can learn to hear it. IME the learning to hear it possibility may be true for a lot of the stuff that people argue about. My advice: stop trying to listen for what you know about. Don't listen for FR or for HD, etc. Just listen to, say, one dac verses another dac without trying to notice anything technical in particular. You might focus on little details of the music, how piano cords form textures. Maybe how cymbals sound, like bursts of noise, like rushing water, like sizzling grease, like a real cymbal struck with a stick with a plastic tip, or maybe with a wood a wood tip, etc. A whole lot of detail is there if one listens and if the reproduction system is good enough.
I might add that a lot of small audible effects are also easy to mask. Listening on an unfamiliar system can be disorienting as well. Adding DBT relays and switches can add masking distortions or not. Make sure the designer of such systems knows how to make sure the the test gear does not mask the effects being tested for.
For anyone who has practiced mindful meditation, there may be an analogy to mindful listening. One trick to achieving a mindful meditation state is to attempt to be consciously be aware of sensation of every part of one's body all at once. its done by starting with the toes working up the legs, and adding awareness of each little part while dropping awareness of none. If done properly the amount of information is overwhelming. It easily fills up conscious awareness. However it takes practice not to drop awareness of some already sensed parts as new parts are added. So it is with listening to music. The brain filters out much of existing but nonessential sensory information automatically, particularly if we are busy thinking about FR, HD, or anything thing else in particular.
Anyway, maybe that's one way to think about the problem of trying to learn to develop a mindful type of listening. It may take some practice filling the mind with attention to a lot of little details of musical sounds, and little details of the sound of a space where music is recorded. On mindful meditation: https://www.frontiersin.org/articles/10.3389/fnhum.2013.00012/full
Last edited:
This is complete nonsense. You would need a very grossly nonlinear system to produce a frequency that's not harmonically related to the input signal. A simple broad resonance would just pass the 4.6 kHz.
I haven't got the equipment at home to do such a test.
Listening tests may be quite objective, but they require some sophistication on listener's part, kind of understanding of what one needs to hear. When some random people are asked to evaluate reproduced sound, most of the time they would hear no difference.
My favorite test sound is applause. I have a SACD of Carlos Kleiber playing Beethoven's 7th Symphony. The first track is 1 minute applause. One should hear not a homogenized noise, but multitude of distinct claps coming from different directions and distances.
Too many systems aren't resolving enough to reveal subtle differences, e.g. between coupling capacitors.
My favorite test sound is applause. I have a SACD of Carlos Kleiber playing Beethoven's 7th Symphony. The first track is 1 minute applause. One should hear not a homogenized noise, but multitude of distinct claps coming from different directions and distances.
Too many systems aren't resolving enough to reveal subtle differences, e.g. between coupling capacitors.
Not only John Curl, but other eminent designers like Lynn Olson or Gary Pimm.
I wanted to provide an update on my comparison of the two Moreplays I have. One Moreplay was built with Audyn + 3.3uF coupling caps with Dueland copper bypass caps. The other Moreplay was initially built with the stock orange drop 2.2uF film coupling caps and no bypass caps and was later upgraded with V-Cap ODAM 2.2uF coupling caps with Dueland copper bypass caps. The unit with the Audyn + caps was well burned in. Tubes are identical and switching tubes between the Moreplays had no noticeable impact on sound. Cabling is slightly different between the units but high quality RCA interconnects and shielded power cords were used on both units. Levels were matched to 0.1 dB using pink noise. I used a remote controlled RCA switch box to do A/B testing and focused on tracks with interesting opening sequences that could be easily and quickly listened to before switching to the other preamp. Conclusions are:
1) Moreplay with Audyn+ and Dueland Bypass caps was slightly better than the Moreplay with stock orange drop film caps after 24 hours of burn in. I didn't burn that unit in for any longer before I decided to install the V-Cap ODAM caps and Dueland bypass caps.
2) After 24 hours of burning in the Moreplay with the ODAMs the sound was nearly indistinguishable from the Moreplay with the Audyn caps but my preference was for the Moreplay with the Audyn caps.
3) After about 160 hours of burn in the Moreplay with the ODAMs was noticeably, but not dramatically, better than the unit with the Audyn caps. The improvement was a sort of solidity or density of the instruments and better resolution and tonality in the upper end so things like cymbals or chimes sounded better with more resolution and also better decay. These differences were minor but real and repeatable across a range of tracks. The chimes and cymbals on the opening of James Newton's Skye illustrated this difference well. https://open.qobuz.com/track/139835273
The differences I heard were subtle, but I think would be noticeable to any experienced listener who had trained themselves to hear the differences between various amps, tubes, and yes cables. Generally, noticing the differences required playing a portion of a track paying attention to specific details and noticing any change between the two preamps. I certainly don't have an adequate audio memory to have permitted me to notice a difference if I was forced to get up and switch cables etc. etc.. Whether any improvement/difference is worth the extra expense of premium caps is really a personal decision. But in my experience the stock Moreplay was improved with premium caps and better NOS tubes.
I appreciate the objectivist view point that these types of components measure identically and consequently have no audible difference but like most listeners who have bothered to take the time to do careful A/B testing I feel that belief is simply wrong. But I'm not interested in restarting that argument and simply wanted to report what I heard listening as closely, and as carefully, as I could.
1) Moreplay with Audyn+ and Dueland Bypass caps was slightly better than the Moreplay with stock orange drop film caps after 24 hours of burn in. I didn't burn that unit in for any longer before I decided to install the V-Cap ODAM caps and Dueland bypass caps.
2) After 24 hours of burning in the Moreplay with the ODAMs the sound was nearly indistinguishable from the Moreplay with the Audyn caps but my preference was for the Moreplay with the Audyn caps.
3) After about 160 hours of burn in the Moreplay with the ODAMs was noticeably, but not dramatically, better than the unit with the Audyn caps. The improvement was a sort of solidity or density of the instruments and better resolution and tonality in the upper end so things like cymbals or chimes sounded better with more resolution and also better decay. These differences were minor but real and repeatable across a range of tracks. The chimes and cymbals on the opening of James Newton's Skye illustrated this difference well. https://open.qobuz.com/track/139835273
The differences I heard were subtle, but I think would be noticeable to any experienced listener who had trained themselves to hear the differences between various amps, tubes, and yes cables. Generally, noticing the differences required playing a portion of a track paying attention to specific details and noticing any change between the two preamps. I certainly don't have an adequate audio memory to have permitted me to notice a difference if I was forced to get up and switch cables etc. etc.. Whether any improvement/difference is worth the extra expense of premium caps is really a personal decision. But in my experience the stock Moreplay was improved with premium caps and better NOS tubes.
I appreciate the objectivist view point that these types of components measure identically and consequently have no audible difference but like most listeners who have bothered to take the time to do careful A/B testing I feel that belief is simply wrong. But I'm not interested in restarting that argument and simply wanted to report what I heard listening as closely, and as carefully, as I could.
Thanks for the detailed and well considered listening experiment. Your listening conditions are very like my own. I'm an ex-pro musician, conservatoire trained, and I'm extremely familiar with the tonality of acoustic instruments. I have selected four or five tracks which are designed to reveal particular tonalities like woodwind, strings, grand piano, vocals, drum kit with particular emphasis on cymbals, and one or two other things. I am extremely familiar with these tracks, which I have been using for years. So very targeted listening with familiar sounds.
I do need some coupling caps right now in the 0.47 - 1uF range, so I'll have to buy something. And at a sensible price, not boutique prices. I can bypass with Russian FT-2 teflon caps if that improves the sound. I was wondering about DC Link caps from Vishay or Kemet or others. I use them for PSU and cathode bypasses but haven't tried them for coupling.
I do need some coupling caps right now in the 0.47 - 1uF range, so I'll have to buy something. And at a sensible price, not boutique prices. I can bypass with Russian FT-2 teflon caps if that improves the sound. I was wondering about DC Link caps from Vishay or Kemet or others. I use them for PSU and cathode bypasses but haven't tried them for coupling.
IME DC Link caps don't sound good. Have some here sitting in a drawer to remind me of that. If you don't want the really good stuff, try to find a reasonably good polypropylene and foil cap. If that's too expensive try metalized polypropylene. The pulse rated Wima metalized polypropylene caps are decent enough, if you can find one of as many uf as you need. If you follow smaller value caps with a FET buffer that can be sufficient, sometimes a bipolar buffer can work too.
To my ear, the FT-2 teflon was not remarkable. I found Russian silver mica caps more interesting as a bypass. If you can find them inexpensively. https://www.ebay.com/itm/254730247967?var=554599704961&hash=item3b4f1b1f1f:g:gRcAAOSw3Zpfu9~U
I love the FT-2 caps, but I know people differ. I also have silver mica caps but i prefer the FT-2s. I can try both.
So still looking for a good value coupling cap in the 4.7-1uF range so suggestions welcome.
What are the Audyn Plus caps? Are they the MKP Plus? HiFi Collective have other Audyns.
https://www.hificollective.co.uk/catalog/audyn-capacitors-mkp-plus-cap.html
So still looking for a good value coupling cap in the 4.7-1uF range so suggestions welcome.
What are the Audyn Plus caps? Are they the MKP Plus? HiFi Collective have other Audyns.
https://www.hificollective.co.uk/catalog/audyn-capacitors-mkp-plus-cap.html
FWIW I tried a few combinations for 1uF coupling caps on triodes. Very small changes. First 4uF Russian PIO bypassed with .1uF silver mica. Then 1uF Obliggato Gold with and without silver mica and then 1uF copper Miflex. To me each step was a little better, but for pure value it is hard to beat the first. Worth noting I have no hearing above 12k.
Back when serengetiplains was an active member I sent him a Russian SSG for testing in exchange for one of his flourinert infused Russian Teflons. In his measurements the silver mica's DF was competitive with the Telfon, i.e. superlative. Interesting to see confirmation the SSG is also mechanically inert. Much though I love 95% of what an SSG does I can't get past a sense they impart a subtle resolution floor related (complete woo guess) to the high DA. Even a Solen or Wima bypassed with 10% of its value in FT-1 Teflons just sounds better to me. Or alleviates my nervosa. The amp has no high DA caps, electrolytic or otherwise, anywhere in circuit so that might be a factor.
I've been following two very long capacitor threads:
https://www.humblehomemadehifi.com/download/Humble Homemade Hifi_Cap-Test-Ratings.pdf
https://www.head-fi.org/threads/orgy-of-capacitors-the-cap-thread.284863/
Should give plenty of subjective information for potential purchasers.
https://www.humblehomemadehifi.com/download/Humble Homemade Hifi_Cap-Test-Ratings.pdf
https://www.head-fi.org/threads/orgy-of-capacitors-the-cap-thread.284863/
Should give plenty of subjective information for potential purchasers.
They are both great value performers. May come down to personal taste. I forgot to mention, with the SSG, try them with the metal tabs removed. I just solder 24 gauge signal wire directly into the little tubes. I believe the metal tabs are silver plated steel. Seems cleaner without them.