Why is high capacitance good?


Hello- I was looking into Cardas speaker cables, and couldn't help to notice as you go up the line---the dramatic increase in capacitance in his speaker cables (450/ft on the top of the line). I believe MIT and those guys crank up capacitance as well.

Is this why cables can sound "warm" (the Cardas mojo)? Most of the reading on the web says capacitance should be at a pretty moderate level.

Cheers,

KeithR
keithr
In itself, high capacitance in a speaker cable is not good. And with some amplifier designs, especially marginal ones, it can lead to poor sound, oscillations, or even damage, unless a Zobel Network is used to compensate.

However, high capacitance can be a side effect of designing the cable to have ultra-low inductance. Low inductance being likely to improve the neutrality of the cable, everything else being equal, although lowering it beyond a certain point will be unnecessary overkill.

High capacitance can also be the intended result of designing the cable to have extremely low characteristic impedance, characteristic impedance being approximately equal to the (square root of (inductance per unit length divided by capacitance per unit length)).

Some speaker cables are marketed based on the theory that their characteristic impedance should approximately match the speaker's impedance. Although I don't question that those cables can and do provide excellent performance in many systems, imho that theory makes little sense, and imho that theory is not what accounts for their performance. I say that for two reasons:

1)For any remotely reasonable cable lengths, characteristic impedance is a concept that is only applicable at frequencies in the radio frequency range, or that at least approach being in the rf range. At those frequencies, the cable's characteristic impedance should closely match the impedance presented by the load, or some of the incoming energy will not be absorbed by the load and will be reflected back toward the source, creating what are known as vswr (voltage standing wave ratio) problems.

2)At the high frequencies at which characteristic impedance may begin to become significant, speaker impedance will be much different than it is in the audio range.

I calculate based on the published capacitance and inductance numbers for the particular cable you referred to that its characteristic impedance is around 5 ohms, right in the vicinity of typical speaker impedances at audio frequencies.

For more conventionally designed cables that figure would typically be in the rough ballpark of 75 or 100 ohms. Which, interestingly, is probably a much better match to the impedance of typical dynamic speakers at frequencies that are in the lower part of the rf range, where the effects of characteristic impedance may start to become significant. Tweeter voice-coil inductance will cause the speaker impedance to rise substantially at those frequencies.

That said, based on user comments many of those cables do give excellent results in many systems. And the combination of inductance which is low to the point of overkill, with capacitance that is unusually high, would certainly seem likely to make them sound DIFFERENT than more conventionally designed cables.

Best regards,
-- Al

To get rid of any negative effects of capacitance, inductance etc - Spectron Audio use "Remote Sense" speaker cables, see - http://spectronaudio.com/cables.htm.

Basically they take full advantage of their ultrafast feedback loop and include speaker cable into it - making amp/speaker the single amplification system.

Does anybody else use the same concept?

Thank you
Al- thanks for the detailed (and fascinating) response. How high are radio frequencies that you mention? Some of these particular cables may actually attenuate high frequencies (MIT with Wilson seems to be a popular combination because of it) so perhaps you are right.

KeithR
Hi Keith,

RF transmission line effects, including characteristic impedance and vswr/reflection effects due to impedance mismatches, become significant when the length of the cable becomes a "significant" fraction of the wavelength of the signal (which is inversely proportional to frequency).

What is "significant" is a matter of degree, of course, and depends on the particular application, and for audio it would seem appropriate to define tolerances more tightly than for most other applications.

But to provide some perspective, the wavelength of a 20,000Hz signal propagating through a wire is in the rough vicinity of 6 miles. The wavelength of a 20Hz signal propagating through a wire is in the rough vicinity of 6,000 miles. The length of a typical speaker cable would certainly seem to be utterly insignificant in relation to those numbers.

The long-wave radio band is commonly thought of as beginning at around 150kHz, so it might be reasonable to consider some point between say 50 and 150kHz as a conservatively drawn point of demarcation between rf and ultrasonic frequencies. Long-wave transmitting antennas, btw, are HUGE, due to the long wavelengths they are transmitting.

DIGITAL audio signals, btw, have frequency components that extend well into the rf range, and rf transmission line issues are therefore very much applicable to connections between transports or other digital sources and dacs.

Best regards,
-- Al
Hi capacitance cables may be used to 'tune' a too bright system. Too bright a system may be fixed in other ways.

I think that's a bad approach. I personally try to find low capacitance cables, when such spec's are known.
Many cable companies don't advertise such things, do they?