Let's discuss caps once more.

[JimRayden]

Here's a few interesting links:
http://members.aol.com/sbench102/caps.html
http://members.aol.com/sbench102/caps2.html

I did a search on the forums and found that many accuse the ceramic caps of being "harsh" sounding. I'm very impressed by the above research because the ceramics look so damn different from all the others. What do that curved line tell me and what's the deal with the ceramic caps?

I'm currently using all ceramic caps whereever I can in my pedals, I just want to know if it's a wise choice or not.

----------
Jimbo

[Fp-www.Tonepad.com]

I use ceramic for capacitances under 1000pF.

Some like to use silver mica for those values, I don't.

Fp

[JimRayden]

Any audible reason for your choices?

Well, the ceramic and the tantalum look like mojo items according to that research.

http://members.aol.com/sbench102/cap4.jpg
http://members.aol.com/sbench102/cap8.jpg

----------
Jimbo

[Fp-www.Tonepad.com]

I just can't hear any difference between the ceramics and silver micas. So ceramics are more convenient given their size/price.

Note: I use this as a general rule given the use of capacitances under 1000pF in the guitar effect circuits we use. There's always the exception and that's when they're used as coupling caps, I won't use a ceramic as a coupling cap because they can be microphonic.

Fp

[brett]

Hi.
Given the test conditions, I'd wonder about the relevance to stompboxes.

The signal level was held constant at about 70 volts RMS at 600 Hz across the capacitors. (for about 26mA signal current).

For me, it's a matter of ceramics for cost for caps under 1000pF (and MKT 1000pF to 1uF).  I've tried polystyrene in some bypass and feedback positions, and not noticed any significant difference.  Silver-mica probably fall into the same category.
I *never* use monolythics, and avoid electrolytics where possible.
cheers

[george]

I have a problem understanding how different dielectrics could cause a "harsh" sound, without some kind of scientific basis for making this judgement, the only differences between capacitors would be the capacitance (and variation within tolerances) and DC leakage.

What does "harsh" mean anyway?  I guess it means "not enough filtering of high frequencies in solid state clipping circuits".

probably what happened was someone replaced a "harsh" ceramic coupling cap with an equivalent mica cap that was in fact actually of higher capacitance and did a better job of filtering.  and a "mojo legend" was born ...

IMHO variations within tolerance would probably swamp any differences due to D-M properties ...

[R.G.]

What do that curved line tell me and what's the deal with the ceramic caps?

That the current through the caps is not a linear function of the voltage across the caps, and vice versa. That is, the cap is distorting the signal, however slightly. This tends to be a problem only near the RC rollover frequency, as other effects swamp the cap's variation away from there.

I'm currently using all ceramic caps whereever I can in my pedals, I just want to know if it's a wise choice or not.

Your ears know. If you like the sound, it's wise. There will always be other people that don't like your pedals.

I just can't hear any difference between the ceramics and silver micas. So ceramics are more convenient given their size/price

.
If you can't hear it then it absolutely does not matter - to you at least. However, once there is any repeatable, observable difference, it's a poor bet to say that no one could ever hear it.

I have a problem understanding how different dielectrics could cause a "harsh" sound, without some kind of scientific basis for making this judgement,

(1) dielectric absorption and nonlinear V-I ; go read the articles Jim refers to.
(2) All ceramic caps are to some degree piezoelectric. That is, they vibrate mechanically with signal, and generate signal with mechanical vibration. That means that what's going on with the cap mechanically will affect the signal through it.

IMHO variations within tolerance would probably swamp any differences due to D-M properties ...

That's a great opinion, and one I held for years - until I met a guy who could reliably tell me what kind of cap I was using no matter how well the capacitance was matched. Never bet on someone else not being able to hear something just because you can't unless you can repeatably demonstrate no signal differences by measurment. The differences exist - however faintly - so it could well be that someone can hear them.

[george]

R.G. you are the coolest.

Just don't start rhapsodising about carbon-comp resistors ... I can't afford them ...   

[Paul Perry (Frostwave)]

One more thing about the non-linearity and/or microphonic nature of ceramics.. there are many different types of ceramic material. The most non-linear are (I believe) the very small (for their value) ones designed for computer board bypass applications. It is said htat you can epoxy those to a screwdrver & use them as a pickup when listening for noises in a car engine.
There was an extensive (multipart) investigation of cap non-linearity in Electronics World (UK) a couple of years ago.

[Fp-www.Tonepad.com]

You also have to remember that there are very few under 1000pF capacitors in stompboxes compared to other values. That's another reason for their small impact in the overall sound, IMHO.

Fp

[WGTP]

I would think location has a lot to do with it.  It the cap is in the big middle of a dual clipping meltdown, it might not be something audible, but at the tail end of a circuit???  Sometimes the word "grainey" is used to describe ceramics.  Sometimes it might be desireable to "liven" things up.  Sometimes you may be going for smooooth.   

R.G.'s point is something I think we forget alot.  People's ears are sensitive to different things and if your over 50 and have been jamming hard for the last 35 years, you may be lucky to hear anything.   

[JimRayden]

So you all try to avoid them huh?

Well the effects I build are grainey enough, they wouldn't mind getting any grainier.

----------
Jimbo

[aron]

I think I've said all I can about ceramics long ago, but it's really true.... since I've switched away from them, my pedals are much smoother and I like the tone a lot better. That being said, lots of older pedals used lots of ceramics and they sounded fine. I just choose not to use them unless I want a certain tone. You can really hear them in amps..... just sub out one or two and you should be able to hear the difference.

I use them only when necessary (small values etc....). OTOH, people hate the green film caps, but I like them fine.

[Eb7+9]

just to make things a little more clear here, cap linearity testing using a sine wave has nothing to do with DA which manifests itself in a transient way - DA's not a steady-state phenomena ... another thing should be very clear, both DA and linearity effects in caps get proportionally worse with signal amplitude - so, where they are placed in a circuit IN TERMS OF SIGNAL LEVELS can very well affect the dynamic transparency of a circuit ... DA is also "loosely" additive ...

case in point, the Fender Tweed Bassman/Twin preamps have a cathode bypassing electrolytic cap in the first gain stage only, where signal amplitudes are at their lowest in the signal chain ... these amps are very hi-fi sounding when operating at low volume ... while in BF Reverb amps the same type of cap is found in the cathode circuits of every gain stage in the preamp ... those amps are NOT very hi-fi sounding when operating at low volume ... take all those caps out and, aside from the accompanying gain reduction, the hi-fi quality shows itself ...

the same reasoning applies to many effect circuits ... Doug's Meteor is a good example (for similar reasons to the Bassman preamp) - it's definitely got more transient transparance than a stock TS808 circuit ... in low gain effects circuits like the Univibe filter stages for another example, you can replace all the signal caps by film and find little diff to using lousy NP or ceramic caps - the signal level there is basically instrument level so it makes little difference what kind of caps you use to pass signal ... and in a FuzzFace the second emitter usually sees a 22uF electro, if you can parallel a bunch of film caps to the same value you will typically hear/feel some difference ... there, the signal transients are large and sharp-edged, and the type of cap plays a bigger role because of the high-transient nature of the signal ...

saying categorically that one kind of cap is good and that another kind is bad without interpreting or relating underlying assumptions to a specific situation is a bit simplistic ...

something else to consider, any non-linear transfer in your circuit - whether it's coming from a gain stage or a passive component - will likely generate added harmonics to your signal ... this may cause your circuit to appear more harmonically interesting to the ear than a dynamically linear circuit ... so there's a two-way tradeoff to be aware of here depending on the signal-amplitude situation // that's why using old-school cap combinations (Ceramic/Electro) often gives a "vintagey" feel to some circuits ... IMO there's a balance of transient response and harmonics that gives that signature ... some like it some don't - it depends on many other things too ...

if you're really pushing this "fi" goal, and this is where it gets a little more complicated, to really flex the transient transparency muscle in analogue circuits there is one remainding proplem to look at - namely, seeing how the transient response of the signal caps interacts with the transient response of the gain stages ... they are, after all, highly coupled phenomena ... so if you're using op-amps or discrete gain stages the DA effects of the coupling caps can play more or less a role in the overall fidelity picture ...

as of yet, there's no scientific theory on codifying transient distortion - we're stuck with our own subjective sense of empericals ... that's why the THD thing in hi-fi audio is now largely considered meaningless whereas a few decades ago it was considered a golden measuring stick ... and it's also why designing for more "fi" still follows a heuristic approach ...

~jc

[brett]

Hi.
While I agree with almost all of the above, but I think that there are a couple of concepts (scientific or otherwise) that generally apply.  Intermodulation distortion (transient and otherwise) is rarely pleasant.  Harmonic distortion is another thing.  If you want your guitar to sound like it's got a sick Dalek inside, get an amp with plenty of intermodulation distortion (or buy a ring modulator and go whacky on purpose). 

Straight transistor distortion always sounds non-harmonic (and bad) to me.
cheers

[Paul Perry (Frostwave)]

in BF Reverb amps the same type of cap is found in the cathode circuits of every gain stage in the preamp ... those amps are NOT very hi-fi sounding when operating at low volume ... take all those caps out and, aside from the accompanying gain reduction, the hi-fi quality shows itself ...

This may be true JC, but, pulling out those caps not only reduces the gain, it puts a lot of negative feedback in. Which might in itself reduce the distortion from other sources.

[R.G.]

Darn it jc, didn't you ever learn to use capital letters and periods? I'm OK with the occasional ellipsis for emphasis and verisimilitude, but it's not the only punctuation symbol we have. It makes your stuff hard to read. There's no shortage of periods and question marks, and the shift keys are easy to use. It's the next best workout for pinkies next to actually playing a guitar. 

The hifi people, especially the tweakos, can laps solidly lyrical over the combination and recombination of components, literally adopting the language of cooking to describe what they think they hear. They talk about using combinations of components and synergetical mixes of parts, even tuning the board material to make the sound somehow different. To the extent that there's something - anything! - that can be measured and has some potential to affect audio frequencies, I'm willing to allow that there are some people who can hear the difference, as I said above. But the idea of using teflon circuit boards and nickel plated circuit traces with silver bearing solder to get some internal clarity that can't be measured is a little silly.

What a lot of the discussions about capacitors miss is that the caps have to matter in a circuit sense before you can hear them at all. If, for instance, you have a simple RC treble cut filter, at frequencies way below the RC time constant, the capacitor does not matter - it's effectively an open circuit, and the resistor dominates the current and voltage that come through. Likewise, at frequencies significantly above the RC rolloff, the capacitor is effectively a short circuit, and the resistor again dominates the signal that gets through. It's only in the frequencies around the RC rolloff that the cap has any significant effect on the signal - and hence the sound - that comes through at all. The capacitor's effect on the signal is swamped by other things at any other frequencies.

as of yet, there's no scientific theory on codifying transient distortion - we're stuck with our own subjective sense of empericals ...

Mr. Otala will be disappointed to hear that.

Interestingly enough, it has and can be shown in repeatable tests that the human ear is remarkably INsensitive to even large amounts of distortion that are transient in nature. How exactly does that play into the mix do you think?

that's why the THD thing in hi-fi audio is now largely considered meaningless whereas a few decades ago it was considered a golden measuring stick ... and it's also why designing for more "fi" still follows a heuristic approach ...

Actually, I don't think that's accurate.
(a) Who thinks it's meaningless?
(b) Who took the survey? What were the numbers?
(c) When was it that THD was a golden measuring stick?

The obvious effects of clipping, gross THD and IMD, and limited frequency response have been pretty well plumbed. I think that the additional distortion effects are just hard to get a handle on, and so the "designers" have to choose whether to work really, really hard to dig out the remaining issues or they can sell products which have a base of oxygen free copper and silver bearing solder in a sauce of teflon circuit board and silver interconnect wire. I think that desiging for more "fi" is getting lost in the snake oil. The level of performance of audio equipment in general has gotten very, very good compared to where it was when the distortion wars were being fought in the 1960s. The way I'd put it is that audio equipment has gotten good enough that we can now hear junk that was always there, but hidden by the gross THD and IMD.

Meaningless? Put the THD and IMD back in, then tell me which amplifier has the silver bearing solder in it. And give a read to Self on Amplifiers.

[mydementia]

Why are monolythics so bad?  They seem to be about the only sort my local surplus store carries...

[Paul Perry (Frostwave)]

Why are monolythics so bad?  They seem to be about the only sort my local surplus store carries...

They are only 'bad' if they make an (unwanted) difference. The type of mono ceramic that might make a difference, are the .1 and .01 that are 'small for their value' that are used as bypss in digital stuff. Because, the particular ceramic used, has a particularly high dielectric value, but is also particularly non-linear & temerature dependant, plus those bypass monos have a pretty wide value fluctuation (users only care that they have AT LEAST the rated cap and low enough series inductance).
My opinion is, don't sweat it, get your FX built, and then once it is going you can try experimenting with replacing caps. The 'wrong' cap isn't going to stop anything working.