[Omega]

Many audio cable designs aim to employ low dielectric shields around the conductor. Why?

Background: The dielectric value is a constant placed into equations describing how the electrostatic field at one point in space will propagate to another point in space. Wikipedia does a decent job here:
Relative static permittivity - Wikipedia, the free encyclopedia

Higher dielectric = higher amount of "charge screening" = less transmission of charge.

My hypothesis: When used as cable insulation, higher dielectrics encourage higher capacitance along the direction of the conductor. Capacitance is frequency-dependent and can cause non-linear attenuation in the cable. Non-linear attenuation should be observable in cables of sufficient length or signal strength.

Questions: Can anyone provide insights, anecdotal experience, or actual data to support (or reject) the hypothesis? I'm trying to understand why audio dogma says low dielectric cable insulators are superior, and get a feel for the magnitude of that effect. Is using an insulator with a dielectric constant of 4.0 twice as bad as a dielectric of 2.0?


If the discussion enters forbidden territory for DIY forum, apologies, I'll withdraw it and forget I asked . But I think it is important to build things with at least a basic understanding of *how* they work...so this should be a relevant question to anyone here building cables or electronics with internal wiring. Really, this is akin to discussions of whether a particular output transistor or opamp affects sound, etc.

To facilitate discussion, here are dielectric values for commonly used cable insulators and other known materials:
water: 78
ice: 4.1
nylon: 3.5 (approx.)
polyvinyl chloride (PVC): 3.0
paper: 3.0-3.5
(butyl) rubber: 2.35
polyethylene: 2.25
vaseline: 2.16 (Ha! love to see this used on a cable!)
Teflon: 2.1
air: 1.00

 

[FallenAngel]

Oil-filled cables also exist, kind of funky. Cotton covered also. I really also like a modified recipe of the VenHaus design using a cotton rope as the center dielectric - nice sounding cables result.

 

[JamesL]

I always wondered as a kid, why there was oil leaking from the speakers
Of course now I know better

 

[nikongod]

when you look at the cable not only as a thing to get signal (and maybe power too depending where it is...) from point A to point B but ALSO as a capacitor shunting signal to ground the importance of dielectric is clear. There are also aspects of microphonics to take into account. An insulator that also prevents microphonics can have a VERY audible effect in some applications.

There are VERY few "pure" devices in electronics. Many things share many characteristics of many other things.

Imagine putting a 55pf cap from output to ground on the output of your source: this is a low-capacitance 1M long cable - they get WORSE. This cap could be an "air" cap, with 2 plates 1mm apart, a "teflon" cap, a cap that uses waxed thread as a spacer, or perhaps a PVC cap (you would not buy a signal cap made of PVC)... While some will point out that it is better not to put this cap here, and are correct in this statement, their commentary is better used to provide a solution to this problem.

This analogy of a cap across the output of the source is problematic because first it is a cap! Assuming you say that all wire dielectric materials sound the same you STILL have this capacitance. When you look at what some capacitance does across the output of a source with a highish output impedance (a tube device for example, some solid-state line-outs on CD players...) you see it creates a low-pass-filter. Even in the case of a source with a "medium" output impedance, the ill effects of a "very capacitive" cable can cause audible effects.

 

[kipman725]

@ 20KHz a 55pF capacitor has an impedance of 145Kohm a modern CD player usually has an output impedance of <1Kohm. So to put that in realistic terms that means that the taking into account no other factors the output of the CD player is 0.68% less than at 0HZ (DC). In my opinion that is not noticeable. Although this situation my be different for turntable as they have higher output impedances (not quite sure how high I have never been interested in something that sounds worse every time you use it).

 

[ericj]

Take a look at Stax's low-capacitance cable on their higher end products.

The thing about the dielectric is, they basically put spacers between the wires.

I've been idly considering making ribbon cables for 'stats and maybe for low-voltage 'phones as well using a loom of some sort to bring off-the-shelf wire together with a fiber jacket.

 

[ashmedai]

Say you have a parallel plate capacitor. The following formula describes the effect of the dielectric constant on the resulting capacitance.

[size=small]C = ε[/size][size=xx-small]r[/size][size=small]C[/size][size=xx-small]vac[/size]

Capacitance is bad if you want to transmit a relatively unaltered analog signal, so you have to compromise between preventing cross-interference and preventing capacitive nonidealities.

Additionally, note that it takes work (in the physical sense) to charge a capacitor, so all this is presumably attenuating the signal since the energy to do so must come from the field inducing it. You could also look at it by noting that the polarization of a dielectric is proportional to its electric susceptibility - polarizing the material takes work, thus consumes energy, thus attenuates the signal.

 

[kipman725]

ok guys can anyone explain how my 30MHz scope probes get away with using what appears to be be PVC as a an insulation material and I have no noticeable increase in attenuation from 100HZ - 10MHZ (limits of my function generator)?

what I'm getting at is the effects all of you are on about are just not applicable to audio frequency signals there effect is so small that are going to need very sensitive test equipment just to measure them and unfortunately the human ear is not a sensitive test instrument.

 

[ashmedai]

Well, I didn't say it was good reasoning...I'm partial to cotton, myself. Teflon is nice because it doesn't have the melting problems PVC does if you're soldering, separate from any electrical properties.

 

[nikongod]

Quote:

Originally Posted by kipman725 /img/forum/go_quote.gif @ 20KHz a 55pF capacitor has an impedance of 145Kohm

55pF is the LOW side of things although there are a couple cables with less capacitance. more often you have MORE
Quote:

Originally Posted by kipman725 /img/forum/go_quote.gif ok guys can anyone explain how my 30MHz scope probes get away with using what appears to be be PVC as a an insulation material and I have no noticeable increase in attenuation from 100HZ - 10MHZ (limits of my function generator)?

how capacitive are your leads?
What is the output impedance of your function gen?
considering these 2 things, are you surprised that you can measure a 10mhz wave?

Are you measuring square or sine waves?
Try adding 10K-ohms (plate impedance of a 6DJ8) between the function gen and the scope and measuring a 10k square wave. capacitors start to look a lot like cables all of a sudden

 

[kipman725]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif 55pF is the LOW side of things although there are a couple cables with less capacitance. more often you have MORE how capacitive are your leads? What is the output impedance of your function gen? considering these 2 things, are you surprised that you can measure a 10mhz wave? Are you measuring square or sine waves? Try adding 10K-ohms (plate impedance of a 6DJ8) between the function gen and the scope and measuring a 10k square wave. capacitors start to look a lot like cables all of a sudden

20pF standard probes, 50ohm tap, 1Mohm input impedance to scope. It's a good job music isn't made of square waves (heh apart from inside DAC's and ADC's where ift goes through a delta sigma modulator), but sine waves then. I guess the 6DJ8 is a tube used on the output of preamp but even with this hypothetical 50pF cable and such a high output impedance the frequency response still extends to >300KHz

 

[nikongod]

Quote:

Originally Posted by kipman725 /img/forum/go_quote.gif 20pF standard probes, 50ohm tap, 1Mohm input impedance to scope. It's a good job music isn't made of square waves (heh apart from inside DAC's and ADC's where ift goes through a delta sigma modulator), but sine waves then.

50 ohm tap... you are trying to see the flaws caused by parallel capacitance in certain conditions while using a nearly best case for it avoiding them.

music isnt made of square waves, but a transient or can approximate the edges of one.

It is quite common to watch a 10Khz square wave go through the "device" to estimate the performance of a device quickly. A 10Khz square wave will show some problems that a sine-wave will never show like faults with transients and ringing.
Quote:

I guess the 6DJ8 is a tube used on the output of preamp but even with this hypothetical 50pF cable and such a high output impedance the frequency response still extends to >300KHz

where did you get this number for 50pf for a cable? is that best case or worst case?

I know where, I said 50pf in a previous post. Both times with the caveat that it is the LOWER END of the range. there are VERY few 1m cables with less capacitance than this. Most cables have more capacitance than this, so calling 50pf hypothetical is indeed correct, it was a guess on the wrong side of things. To be sure if i said a higher number someone would have quoted that there are a couple commercial cables in the 40pf range for 1M and that other cables extend as high as 190pf (and more for some designs) for 1m.

measure 1m of canare starquad 4E6S (both blue together, both white together) driven with a 10k hz square wave from a 10k ohm source.

 

[ashmedai]

Um, yeah...take those off the scope and go find a good L/C meter. I think the effect is still very minor...but it sounds like you're measuring it wrong.

 

[kipman725]

I'm not mesuring anything I have been doing calculations on the figures others have been posting. BTW a 190pF cable gives a corner freq of 84KHz still from a 10K output impedance. I'm still not seeing a need for special low capacitance cable even in such extreame circumstances. I supose if you were playing SACD in theory you can have audio infromation (at low resolution) upto 100KHz but almost all players LP filter out anything greater than 40KHz (preumably to improve their THD+N spec as the noise is shaped so that most of it is in the higher frequancies). nikon god what kind of music has square waves in (apart from metalicas new album, thats another story!) as I am interested now and want to do some fourier transforms on it to see if there are any components >80Khz! if so you may have a basis for your arguments (but still unfortunatly only with valve preamps and other high impedance sources).