Why are we going 300 or more directions?


Funny, if you design a hi-speed coaxial cable, the fundamental design is the same. I do mean the same. Physics have honed the basic construct to the same physical design no matter who makes it.

Yes, conductor and shield materials will change some based on the frequency range, but not the design. If you change the basic design, you get worse performance, and not just differentiation (unless worse is your differentiation).

Then we have audio cables. True, audio exist in a frequency range where stuff does change with respect to frequency (impedance drops markedly as frequency is swept from 20 to 20 kHz) but this still doesn't allow willy-nilly designs from A to Z to hold the best electrical ideal.

If there are X number of speaker cable makers, only a small few can be the most ideally right according to physics for audio transmission. What we have is so much differentiation that it is almost humorous.

If "we", as in speaker cable designers, all got in a big room with the door locked and could not be let out till we balanced the design to best effect...what would that cable look like? Why?

Go to any web site and you can't get one-third of the way through before vendors allow misconception to be believed (references to velocity of propagation for instance) that are meaningless in their feint of hand or simply unprovable as to their effect...simply fear you don't have it. For instance, high velocity of propagation allows you to simply lower capacitance, the speed is there, but irrelevant at audio and cable lengths that you use. The signal travels too fast to matter. Signal delay is in the 16ns range in ten feet. Yes, that's 16 trillionths of a second. It's the capacitance folks, not the velocity that you engineer to. But velocity "sounds" exciting.

Audio cable over the years should be under CONSOLIDATION of principals and getting MORE like one another, not less so. I don't see a glimmer of this at all.

The laws of physics say there is a most correct way to move a electrical signal, like it or not. Electrical and magnetic fields have no marketing departments, they just want to move from A to B with as little energy lost as possible. You have to reach a best balance of variables. Yes, audio is a balance as it is in an electromagnetic transition region I mentioned earlier, but it STILL adheres to fundamental principals that can be weighed in importance and designed around.

A good cable does not need "trust me" engineering. An no, the same R, L and C in two cables don't make them the same. We all know it isn't that simple. BUT, the attributes (skin effect and phase responses) that DO make those same R, L and C cables different aren't magic, either.

I've listened to MANY cables this past six months, and it no longer amazes me which ones sound the best. I look at the several tenets that shape the sound and the designs that do this the most faithfully always come out on top.

DESIGN is first. Management of R, L, C, Skin effect and phase. Anyone cam stuff expensive material in a cable, few can DESIGN the right electrical relationships inside the cable. Why be stuck with excessive capacitance (over 50 pF/foot) to get low inductance (less than 0.100 uH/foot) when it's NOT required, for instance. A good design can give you BOTH!

MATERIALS are a distant second to sound quality. They contribute maybe 2 tenths of the total sonic equation in a quality design and ZERO in a bad design. A good design with standard tough pitch copper will exceed a bad design with single crystal cryogenic OFC silver-plated copper. You can't fake good cable design and the physics say so. Anyone can buy materials, so few can do design.

Being different to be different isn't a positive attribute in audio cables. Except for all but ONE ideal design it’s just a mistake.

I've listened to the same cables with dynamic speakers and electrostatic speakers, and the SAME cables always come through with the same characteristics. Good stays good. True, the magnitude of character is different, but the order hasn't moved.

I'm not real proud of the cable industry in general. True transmission accomplishments should reach common ground on explainable principals and that SHOULD drive DESIGN to a better ideal. But, we people do have emotions and marketing.

What do I look for in a speaker cable?

1.0 Low capacitance. Less than 50 Pf / foot to avoid amplifier issues and phase response from first order filter effects where the phase is changing well before the high-end is attenuated. The voltage rise time issue isn't the main reason low capacitance is nice, it's that low capacitance removes the phase shift to inaudible frequencies and doesn't kill amplifiers.

2.0 Low inductance as we are moving lots of CURRENT to speakers. Less than 0.1UH /foot is what you want to see. Good designs can do low cap and low inductance, both.

3.0 Low resistance to avoid the speaker cables influencing the speakers response. The cable becomes part of the crossover network if the resistance is too high. For ten-foot runs, look for 14 AWG to maybe 10 AWG. Bigger isn't better as it makes skin depth management issue too hard to well, manage.

4.0 Audio has a skin depth of 18-mils. This is where the current in the wire center is 37% of that on the surface. The current gradients can be vastly improved with smaller wire (current closer to the same everywhere). How small? My general rule is about a 24 AWG wire as this drop the current gradient differential across the audio spectrum to a value much less than 37%. Yes, that's several wires. Don't go overboard, though. Too much wire is a capacitance nightmare. Get the resistance job done then STOP at that wire count.

5.0 Conductor management. Yes, point four above says more than one wire, many more! And, if you use 24 AWG wire for skin depth management, it can be SOLID to avoid long term oxidation issues. I've taken apart some old wires and it can look pretty bad inside! Each wire needs it's own insulation.

6.0 Symmetrical design. Both legs are identical in physical designs allows much easier management of electricals.

7.0 Proper B and E field management is indirectly taken care of by inductance and capacitance values. The physics say you did it, or you didn't. BUT, you can design in passive RF cancellation if you use a good design, too. Low inductance says that emissions will be low, however, as less of the energy is generating an electric and magnetic field around the wire, thus limiting EMI / RFI emissions.

8.0 Copper quality is finally on the list. It doesn't matter without one to seven! The smaller the wires (infinitely small), the LESS the silver plate will warp the sonics. If the current density is the SAME at all frequencies, then all frequencies see the same benefit. If a wire is infinitely big than the high frequencies will see the majority of the benefit. 20 Hz and 20kHz are at the same current density on the wire surface. But, the gradient difference is too small to matter with 24 AWG wires. If you want silver, let the silver benefit everywhere!

9.0 Dielectrics. Dead last. Why? Because capacitance is driven by your dielectric. If you have the low cap, you have the right dielectric for the design. You HEAR the capacitance and NOT the dielectric per say. True, Teflon allows a lower capacitance for the same distance between wires, thus making lower capacitance. But, if you FOAM HDPE from 2.25 down to 2.1 dielectric constant, it can meet the same cap at the same wall and sound just as good. Careful though, it is now more fragile! It's a trade-off in durability, not sound quality. Teflon isn’t magic. It is expensive.

10.0 This is not last per say as it is CHOICE in design. I do not like fragile cables laying on the floor to be stepped on. Some do. A good cable design should be durable enough to take that late night trip to the TV set with the light low, and then step on your cable by accident. The cable should be user friendly.

Everything above can be calculated by known physics equations with the exception of copper quality on sound. I'll have to hear this on two IDENTICAL cables except wire quality. But, why would a vendor allow you to do that when they can scare you into a more expensive copper? I'll be glad to pony-up if I'm allowed to make the judgement for myself. Or, let be buy it at a reasonable price!
rower30

Showing 3 responses by almarg

JN, thanks very much for the additional insights! All of which strike me as being highly plausible, notwithstanding the fact that I have never before seen them so stated. And I strongly second the subsequent comment by Corazon.

I would commend to everyone else's attention the related comments you provided in the 2005 threads here and here.

So if I understand your comment above correctly, skin effect is an even less important factor than even skeptics of the significance of that effect (myself included) may previously have envisioned. While the storage and delayed release of energy resulting from cable inductance and capacitance (as distinguished from the effects of "dielectric absorption" that are often referred to in cable marketing literature) are likely to be of greater significance than is generally realized.

Interestingly, I recall that in past threads Ralph Karsten of Atma-Sphere, who as you may be aware frequently posts here under the screen-name "Atmasphere," has mentioned that many years ago various experiments he had performed indicated a correlation between sound quality and the closeness of the match between cable characteristic impedance and speaker impedance. That, of course, does not seem explicable on the basis of RF transmission line and reflection effects, and I had suggested in one of those threads the possibility that what he observed may simply have been the result of the low cable inductance that usually goes hand in hand with low characteristic impedance. I said that in part because the impedance of a dynamic speaker at the ultrasonic and RF frequencies for which reflection effects may become significant is usually much higher than it is at audio frequencies, primarily as a result of the inductance of the tweeter(s). This statement in one of your 2005 posts seems to reconcile it all elegantly:
As it turns out, the point of minimal line storage occurs when the characteristic impedance of the cable matches the load..so, in theory, an 8 ohm speaker would work best with an 8 ohm cable impedance..this of course, is not because of reflections per se, but rather, just simply from the calculations of inductive and capacitive storage...
So the bottom line, if I understand correctly, would seem to be that in the case of a speaker cable, and assuming that neutral behavior is desired, minimization of both inductance and characteristic impedance is desirable, up to the point at which characteristic impedance matches speaker impedance (at audio frequencies), with capacitance also being minimized to the extent that it is practicable to do so without significantly conflicting with those goals.

Although, of course, whether or not truly neutral behavior will be subjectively perceived as optimal in a given system by a given listener will always remain an open question.

Thanks again for the excellent and rarely stated inputs!

Best regards,
-- Al
03-17-13: Gregm
Rower30: your original post & further "articles" are very interesting and, if i may say so, a significant contribution to the community.
I'll second that. Your writeups make a great many excellent and important points. Well done!

The one area I would question, though, at least with respect to situations where the wire is not plated, is your emphasis on the importance of minimizing skin effect.
Audio has a skin depth of 18-mils. This is where the current in the wire center is 37% of that on the surface. The current gradients can be vastly improved with smaller wire (current closer to the same everywhere).
First, I would rephrase the second sentence to say that "this is where the current in the wire center is 37% of the TOTAL current, the other 63% flowing in the part of the wire's cross-section that is between the surface and the skin depth."

As you undoubtedly realize, with respect to sub-RF frequencies what skin effect basically does is to increase the resistance of the cable as frequency increases. It can be calculated that under typical circumstances (e.g., cable lengths on the order of 10 feet or so, and used with dynamic speakers, which generally have impedances that are high in the upper treble region and above), the effects of that resistance rise will be a rolloff at 20 kHz of a very small fraction of 1 db. That is likely to be completely swamped by room effects, speaker inaccuracies, inaccuracies elsewhere in the system, and the high frequency rolloff and finite resolution of the listener's hearing.

I can provide such a calculation, if desired.

On the other hand, in the less common situation where the length of the speaker cables is particularly long AND the impedance of the speakers descends to very low values (e.g. 1 ohm) at 20 kHz, as it does in the case of many electrostatics, I would agree that skin effect might conceivably become a marginally perceptible factor. And as you indicated, additional considerations come into play in the case of plated wire.

Also, a minor correction to a statement in your initial post, which was most likely just an oversight. 16 ns is 16 billionths of a second, not trillionths.

Again, though, my compliments on what IMO are a great many excellent and important points that are made in your writeups.

Regards,
-- Al
JNeutron, thanks for your obviously very knowledgeable inputs. Yes, there are clearly a number of oversimplifications and "inaccuracies" (which I prefer to think of as "rough order of magnitude approximations") in what has been said in this thread. However, I would question whether or not the "inaccuracies" are relevant to the underlying points that have been made.

For example, concerning your point about propagation delays I am aware that delays at deep bass frequencies can be far longer than the 1.6 ns/ft ballpark figure cited by the OP. At 20 Hz, for instance, I have seen data indicating that for many cable types delays may be 50 times or more greater than that figure. However, that would still equate to more than 2000 miles per second, which clearly corresponds to a negligible amount of time in the context of speaker cables in a home audio system.

Based on the following comment you made in this 2005 thread, it appears that you would agree with that:
I do agree that the term "prop speed" is rather confusing, as most seem to think it means that the transit time from the amp to the load is of any consideration...IT ISN'T [emphasis added]. But, the term prop speed is directly related to the DC [dielectric constant], L[inductance], and C[capacitance].
I certainly don't disagree with that statement, and I doubt that the OP or most of the others who have responded above would disagree. Nor, I suspect, would they disagree with the following statement you made:
But honestly, it's all in the R, L, C, and Q of the cable. Unfortunately, there's been so much mis-information spread around that high end audio guys end up guessing and trial and error, without much in the way of science..and, realistic measurement of matched Z cables is impossible for most wire vendors, as inductance measurements at the tens of nanohenry level are very difficult to do correctly.
Regarding your points about skin effect, can you provide us with a QUANTITATIVE feel for the degree to which the result provided by an exponential approximation would deviate from the Bessel function result? For example, what would be the difference between the 63% skin depths at 20 kHz for solid core copper wire as calculated based on the two functions?

Finally, a question I ask out of sincere curiosity: Do you design or manufacture audio cables, and if so, which ones?

Regards,
-- Al