Do electrolytic caps have a "tone"?

[Esppse]

Hey,

I noticed in most circuits, electrolytics are not in the audio path, and I was curious if there would be a tone difference between higher quality components, and the cheaper tsyda stuff.

Now I understand there is a debate over regular caps, and yes there is a difference in my opinion with same value, different material, because I took 5 caps of same value with different material like ceramic, film, pio etc... and socketed them in the same spot, there was unbelievably a very noticeable shift in sound. So imagine all the caps being different and the possibility of combinations.

I currently don't have access to a bunch of different electrolytic brands and sizes, so I was wondering if anyone knows or have done the socketing experiments I have. I can imagine not being in the audio path could have less effect? The test I did before was in a filtering cap position, so the roll off and color was very easy to notice.

[idy]

When you did you comparison did you measure them to check their actual as opposed to marked values?

[Fancy Lime]

No, caps do not have a "tone", only (electrical and other) properties. Same goes for all other components. When people hear differences between cap materials, most of the time they actually hear differences in capacitance value. The important thing to understand are tolerances. Some cap materials, like film or mica and some ceramic types (COG or NP0) have very tight tolerances of a few percent and little dependence on temperature or voltage across the cap. That is why these are preferred for any and all applications, where the exact value matters. Most ceramics, electrolytics or tantalums have much higher tolerances, up to 50 percent. So a cap that claims to be 100nF may indeed be 50nF or 150nF. You would have to measure it to know. These types also change capacity to a larger degree depending on temperature and, more importantly for us, voltage applied across them. This means the capacitance of the cap will change with the signal amplitude, sometimes quite a lot. In an audio signal, we might notice this as distortion if the cap is in a critical filter. However, compared to the "the values are different because of tolerances" effect, the distortion is usually negligible.

So if you really want to know how different various cap materials sound, my advice would be to measure the real capacitance of the caps in the circuit that you are using them in under power, make sure they are all exactly the same (within 5% or so, better is not realistically possible without going through hundreds of caps) and perform a blind test. Record each cap ten times and have someone play all recordings back to you in random order. Try hearing which recordings of the same material are from the same or from different caps. I would be rather surprised if there was actually any audible difference.

TL;DR: Components have tolerances. If the exact value matters (like in a filter), use a low-tolerance type like film, mica or C0G/NP0. If it doesn't, don't bother.

Cheers,
Andy

[Esppse]

Hmm, I did not measure the capacitance in my test. I have recently got a capacitance meter. So measuring them by themselves is different than in circuit, even a small 9v circuit? How would I measure them when it's in a Pedal?

I will try that test again with the socket.

And as far as audio grade electrolytic, is it just a tighter tolerance or is the cap construction going to be different? I'm trying to figure out how to spot a quality electrolytic because they all look the same, and if it will actually last longer.

[andy-h-h]

I'm trying to figure out how to spot a quality electrolytic because they all look the same, and if it will actually last longer.

Just read the data sheet or product specs as there will be tips there for you.   Tolerance, temperature range, hours rated for etc.  Look at derating for a better lifespan as well.  i.e.  use a higher spec than needed, but not excessively so as that just costs too much and can be too big.   A common example is using caps rated to at least double the voltage they are expected to see in service.   

[amptramp]

If you look at manufacturer's data for electrolytic capacitors, they usually present an impedance vs. frequency diagram that shows the impedance declining with increasing frequency until you get to a point (usually within the audio band) where the impedance starts rising with frequency due to the spiral winding of the foil and the reluctance of the dielectric to behave at higher frequencies.  In high-quality audiophile capacitors, you also get pure copper leads rather than tin over copper plating over steel leads and this can be shown by lifting a capacitor with a magnetized screwdriver or a bigger magnet for a bigger capacitor.  Audiophile capacitors may have the impedance drop down lower and the resonant point (where the impedance goes from declining to rising with rising frequency) is moved higher up in the audio spectrum.  The audio world used to recommend Black Gate capacitors made by Rubycon but the last ones were made in 2006 and not many people trust fifteen-year-old capacitors now.

[Ice-9]

No, caps do not have a "tone", only (electrical and other) properties. Same goes for all other components. When people hear differences between cap materials, most of the time they actually hear differences in capacitance value. The important thing to understand are tolerances. Some cap materials, like film or mica and some ceramic types (COG or NP0) have very tight tolerances of a few percent and little dependence on temperature or voltage across the cap. That is why these are preferred for any and all applications, where the exact value matters. Most ceramics, electrolytics or tantalums have much higher tolerances, up to 50 percent. So a cap that claims to be 100nF may indeed be 50nF or 150nF. You would have to measure it to know. These types also change capacity to a larger degree depending on temperature and, more importantly for us, voltage applied across them. This means the capacitance of the cap will change with the signal amplitude, sometimes quite a lot. In an audio signal, we might notice this as distortion if the cap is in a critical filter. However, compared to the "the values are different because of tolerances" effect, the distortion is usually negligible.

So if you really want to know how different various cap materials sound, my advice would be to measure the real capacitance of the caps in the circuit that you are using them in under power, make sure they are all exactly the same (within 5% or so, better is not realistically possible without going through hundreds of caps) and perform a blind test. Record each cap ten times and have someone play all recordings back to you in random order. Try hearing which recordings of the same material are from the same or from different caps. I would be rather surprised if there was actually any audible difference.

TL;DR: Components have tolerances. If the exact value matters (like in a filter), use a low-tolerance type like film, mica or C0G/NP0. If it doesn't, don't bother.

Cheers,
Andy

Nice informative information on why tolerances can make a difference over cap type, but saying electrolytics are 50% tolerance if a bit misleading. It certainly is more difficult to get a tighter tolerance in a capacitor than it is in a resistor but all the caps I use whatever the type are always a max of 5% or better.

[Fancy Lime]

No, caps do not have a "tone", only (electrical and other) properties. Same goes for all other components. When people hear differences between cap materials, most of the time they actually hear differences in capacitance value. The important thing to understand are tolerances. Some cap materials, like film or mica and some ceramic types (COG or NP0) have very tight tolerances of a few percent and little dependence on temperature or voltage across the cap. That is why these are preferred for any and all applications, where the exact value matters. Most ceramics, electrolytics or tantalums have much higher tolerances, up to 50 percent. So a cap that claims to be 100nF may indeed be 50nF or 150nF. You would have to measure it to know. These types also change capacity to a larger degree depending on temperature and, more importantly for us, voltage applied across them. This means the capacitance of the cap will change with the signal amplitude, sometimes quite a lot. In an audio signal, we might notice this as distortion if the cap is in a critical filter. However, compared to the "the values are different because of tolerances" effect, the distortion is usually negligible.

So if you really want to know how different various cap materials sound, my advice would be to measure the real capacitance of the caps in the circuit that you are using them in under power, make sure they are all exactly the same (within 5% or so, better is not realistically possible without going through hundreds of caps) and perform a blind test. Record each cap ten times and have someone play all recordings back to you in random order. Try hearing which recordings of the same material are from the same or from different caps. I would be rather surprised if there was actually any audible difference.

TL;DR: Components have tolerances. If the exact value matters (like in a filter), use a low-tolerance type like film, mica or C0G/NP0. If it doesn't, don't bother.

Cheers,
Andy

Nice informative information on why tolerances can make a difference over cap type, but saying electrolytics are 50% tolerance if a bit misleading. It certainly is more difficult to get a tighter tolerance in a capacitor than it is in a resistor but all the caps I use whatever the type are always a max of 5% or better.

Well yes, one thing I forgot to mention, is what "audio grade electrolytic" means. Or at least seems to mean to most manufacturers from what I can tell from the datasheets. The most important thing here, is that audio grade electrolytics usually have much much better tolerances. Probably not because they are produced with tighter tolerances, which is quite difficult, but simply by being selected, which is cheap and easy these days. They are also usually low series resistance types. So all of the negative points I listed in the last post for electrolytics apply when you buy just any electrolytic cap, but to a much lesser degree to audio grade electrolytics. However, it is not always easy to evaluate whether something is actually audio grade or if the seller just claims it to be, especially when buying online. My solution is to just buy cheap electrolytic caps and use them where the exact value or series resistance doesn't matter, unless there really is no sensible way around a good electrolytic. But that is quite rare.

Andy

[vigilante397]

all the caps I use whatever the type are always a max of 5% or better.

50% does seem like a bit of an exaggeration, but I see (and use) a fair amount of +20%/-10% caps for larger values.

[Ice-9]

No, caps do not have a "tone", only (electrical and other) properties. Same goes for all other components. When people hear differences between cap materials, most of the time they actually hear differences in capacitance value. The important thing to understand are tolerances. Some cap materials, like film or mica and some ceramic types (COG or NP0) have very tight tolerances of a few percent and little dependence on temperature or voltage across the cap. That is why these are preferred for any and all applications, where the exact value matters. Most ceramics, electrolytics or tantalums have much higher tolerances, up to 50 percent. So a cap that claims to be 100nF may indeed be 50nF or 150nF. You would have to measure it to know. These types also change capacity to a larger degree depending on temperature and, more importantly for us, voltage applied across them. This means the capacitance of the cap will change with the signal amplitude, sometimes quite a lot. In an audio signal, we might notice this as distortion if the cap is in a critical filter. However, compared to the "the values are different because of tolerances" effect, the distortion is usually negligible.

So if you really want to know how different various cap materials sound, my advice would be to measure the real capacitance of the caps in the circuit that you are using them in under power, make sure they are all exactly the same (within 5% or so, better is not realistically possible without going through hundreds of caps) and perform a blind test. Record each cap ten times and have someone play all recordings back to you in random order. Try hearing which recordings of the same material are from the same or from different caps. I would be rather surprised if there was actually any audible difference.

TL;DR: Components have tolerances. If the exact value matters (like in a filter), use a low-tolerance type like film, mica or C0G/NP0. If it doesn't, don't bother.

Cheers,
Andy

Nice informative information on why tolerances can make a difference over cap type, but saying electrolytics are 50% tolerance if a bit misleading. It certainly is more difficult to get a tighter tolerance in a capacitor than it is in a resistor but all the caps I use whatever the type are always a max of 5% or better.

Well yes, one thing I forgot to mention, is what "audio grade electrolytic" means. Or at least seems to mean to most manufacturers from what I can tell from the datasheets. The most important thing here, is that audio grade electrolytics usually have much much better tolerances. Probably not because they are produced with tighter tolerances, which is quite difficult, but simply by being selected, which is cheap and easy these days. They are also usually low series resistance types. So all of the negative points I listed in the last post for electrolytics apply when you buy just any electrolytic cap, but to a much lesser degree to audio grade electrolytics. However, it is not always easy to evaluate whether something is actually audio grade or if the seller just claims it to be, especially when buying online. My solution is to just buy cheap electrolytic caps and use them where the exact value or series resistance doesn't matter, unless there really is no sensible way around a good electrolytic. But that is quite rare.

Andy

You should always select components to do the job you want them to do, this is why datasheets are so valuable. Classing a component as 'Audio Grade' is pretty much a marketing device not really a genuine specification. Just like how audiophiles want to buy a mains cable for £250 etc. Low ESR can be important in caps as is % tolerance in all components. I think what I am trying to say is that there is a lot of smoke and mirrors around people selling components under than banner of Audio Grade.