The 'Skin Effect' for speaker wire.

[BrianHG]

(Please, no Audiophoolery BS...  I'm only interested in hard true physics and numerical facts...)

I know about the 'Skin Effect' in AC power transmission lines even at 60hz which isn't there when transmitting DC.  I also know the skin effects get worse with higher frequencies.  I was wondering about the extent of the effect with a speaker wire operating at 10Hz to 20KHz.  Would a thick solid core copper wire performs better, or a thick cable constructed of thin stranded copper, or thin isolated solid core strands in parallel bet the best (think of something like a network cable where at each end, all the conductors are twisted together making 1 high current cable).  In my setup, all 3 setups would have the same overall copper thickness as if it were the single solid core creating the same final DC impedance.

Oh, and for those who don't know: https://en.wikipedia.org/wiki/Skin_effect

[wraper]

I would not bother about skin effect. Only tiny part of total power is at high frequencies, and max frequency is not that high anyway. Thus effect is negligent.

[BrianHG]

I would not bother about skin effect. Only tiny part of total power is at high frequencies, and max frequency is not that high anyway. Thus effect is negligent.

Understood, however, this is an exercise in a real effect which actually exists and can be measured and what in theory is the best type of wiring for a speaker given that all 3 wires are identically priced.

[wraper]

I would not bother about skin effect. Only tiny part of total power is at high frequencies, and max frequency is not that high anyway. Thus effect is negligent.

Understood, however, this is an exercise in a real effect which actually exists and can be measured and what in theory is the best type of wiring for a speaker given that all 3 wires are identically priced.

Wire with separate insulated strands (Litz wire) is the best but total overkill and won't make any real difference. Solid wire is the worst. I would use usual multi strand copper wire. Solid wire should not be used for anything other than permanent installations (like mains wiring in walls) anyway.

[RandallMcRee]

(Please, no Audiophoolery BS...  I'm only interested in hard true physics and numerical facts...)

I know about the 'Skin Effect' ...something like a network cable where at each end, all the conductors are twisted together making 1 high current cable....

Oh, and for those who don't know: https://en.wikipedia.org/wiki/Skin_effect

Yep, lookup litz wire. It is used at high frequencies for this reason.

[2N3055]

Skin depth in copper at 50 Hz is 9,2 mm. At 10kHz is about 0,65 mm...

No need for litz wire (it is used at many MHz), but it illustrates the principle.
 
Simple stranded good quality copper cable is perfect.. Individual strands should be 0,2-0,4 mm.
That will take nicely through audio range and comfortably more with very little frequency influence..
Make sure you have decent quadrature,but there is little to be had over 2,5mm² for short wires and smaller power..
If wiring is longer and/or higher power, you can go to 4mm².

[BrianHG]

I would not bother about skin effect. Only tiny part of total power is at high frequencies, and max frequency is not that high anyway. Thus effect is negligent.

Understood, however, this is an exercise in a real effect which actually exists and can be measured and what in theory is the best type of wiring for a speaker given that all 3 wires are identically priced.

Wire with separate insulated strands (Litz wire) is the best but total overkill and won't make any real difference. Solid wire is the worst. I would use usual multi strand copper wire. Solid wire should not be used for anything other than permanent installations (like mains wiring in walls) anyway.

Thanks.  This is what I figured.
Now, is there a further advantage to having all those strands isolated like in my network cable example, or like a normal twisted together wire?

[shakalnokturn]

the single solid core creating the same final DC impedance.

Ouch! We must be in the VLF range here...

I actually wanted to do a Litz vs multi-strand comparison, never got round to it in the end so I can't swear it sounds better.

If there is no noticeable benefit on skin effect at audio frequencies, I can only think of the disadvantage of one conductor breaking in the Litz having much more impact on resistance than similar case with multi-strand.

[helius]

The one place where Litz wire is used for audio is in turntables' tone arms. Not because of the skin effect, but because it is much more flexible and won't work-harden over time.

[ajb]

At 20KHz the skin depth is roughly equal to the radius of 18AWG wire.  Are you planning to drive your tweeters at more than 16A_RMS?

[Gr8fulFox]

You don't need to worry about the 'skin effect' with speaker wire. In my experience, the best speaker wire is old vacuum-cleaner cord; it's long, plenty thick for the purpose, and is as simple to obtain as cutting the cord off of a discarded vacuum cleaner someone left on the curb.

[Fungus]

In my experience, the best speaker wire is old vacuum-cleaner cord; it's long, plenty thick for the purpose, and is as simple to obtain as cutting the cord off of a discarded vacuum cleaner someone left on the curb.

And any "golden ears" who claims otherwise should be busy applying for the James Randi million dollar prize instead of posting here.

[dzseki]

One can also buy silver clad copper cables, those usualy come with PTFE jacket. It is a really cheap experiment to compare to similarly sized copper wires, whether skin effect makes different or not.

[Zero999]

One can also buy silver clad copper cables, those usualy come with PTFE jacket. It is a really cheap experiment to compare to similarly sized copper wires, whether skin effect makes different or not.

That won't make any difference with audio, as the copper underneath the silver plate will still carry most of the current, even at the high end of the audio band.

[dzseki]

One can also buy silver clad copper cables, those usualy come with PTFE jacket. It is a really cheap experiment to compare to similarly sized copper wires, whether skin effect makes different or not.

That won't make any difference with audio, as the copper underneath the silver plate will still carry most of the current, even at the high end of the audio band.

Just like Litz, I just had an idea for a cheap experiment if someone wants to try it.

I vaguely recall that I saw at some point a "speaker wire" made of flex PCB, where the two conductors were like 2" wide traces, while conductor quadrature were small but the surface was large.

[Fungus]

One can also buy silver clad copper cables

There's people selling them, sure. Just like everything else the audio world can invent.

Will they make a difference? No.

[Zero999]

One can also buy silver clad copper cables, those usualy come with PTFE jacket. It is a really cheap experiment to compare to similarly sized copper wires, whether skin effect makes different or not.

That won't make any difference with audio, as the copper underneath the silver plate will still carry most of the current, even at the high end of the audio band.

Just like Litz, I just had an idea for a cheap experiment if someone wants to try it.

I vaguely recall that I saw at some point a "speaker wire" made of flex PCB, where the two conductors were like 2" wide traces, while conductor quadrature were small but the surface was large.

Silver plated copper cable will not be cheap, especially if it's marketed for audio use.

Also how does one objectively test it? The listener needs to be unaware of which cable is connected when performing the test, otherwise cognitive bias will creep in: if they expect the silver to sound better for one reason, then they'll believe it.

[woodchips]

As Bob Pease has pointed out on several occasions, the best speaker cable is ribbon cable with alternate cores as signal and ground.

[Kleinstein]

The ribbon cable has some positive aspects, possibly even an characteristic impedance that could match the speaker  . However termination at the ends can be troublesome. As the speaker cables are usually short compared to the electrical wavelength no need to worry about impedance. The best point may be one can hide it well under the carpet.

Non isolated litz wires don't help much against skin effect, as the current would still try to stay on the outside. The RF litz has isolated strands and the wires are not just twisted but go inside and out. For audio it does not make a significant difference. It's only a tiny fraction of the still small loss in the cable.

The typical twisted pair cable can be a little on the thin side. Otherwise separated cables in parallel would reduce skin-effect.

[wraper]

Non isolated litz wires don't help much against skin effect, as the current would still try to stay on the outside. The RF litz has isolated strands and the wires are not just twisted but go inside and out. For audio it does not make a significant difference. It's only a tiny fraction of the still small loss in the cable.

Have some read https://engineering.dartmouth.edu/inductor/papers/stranded.pdf

[SiliconWizard]

To get an idea of the possible impact of the skin effect, you can use this: http://www.chemandy.com/calculators/round-wire-ac-resistance-calculator.htm

You get to see that the impact of frequency on the wire resistance in the audio range is pretty negligible, unless you are dealing with VERY long cables (like over severals tens of meters) or very high current (and thus a large diameter wire).

Another factor you could consider is the inductance of the wire. Again to get an idea: http://www.chemandy.com/calculators/round-wire-inductance-calculator.htm
And again, unless you are dealing with very long cables, the impact of the inductance is negligible in the audio range (except maybe if you use *very* low impedance speakers - think of it roughly as a low-pass filter).

[cdev]

Flexibility as far as the wire goes, soft copper that's flexible, with good solid terminations, whatever they may be, gets my vote. The less jacket and more wire there is inside the better.

You don't need to worry about the 'skin effect' with speaker wire. In my experience, the best speaker wire is old vacuum-cleaner cord; it's long, plenty thick for the purpose, and is as simple to obtain as cutting the cord off of a discarded vacuum cleaner someone left on the curb.

[Fungus]

Silver plated copper cable will not be cheap, especially if it's marketed for audio use.

Yep.

if they expect the silver to sound better for one reason, then they'll believe it.

It actually does sound better. To them.

[dzseki]

Sorry, I never said there will be difference, it was just a reply to Shakanokturn's post.
And also did not imply the cabke should be "audio grade", in fact I was thinking about (industrial) cables from Belden or Huber+Suhner. A few years ago I  stumbled into a pool of 15m such cable in a local shop for something like 8€, sure enough I bought it. Also I have to admit that I use Silver clad copper-PTFE RG-179 coaxial cables for audio interconnect, but simply because they are the most rugged for the money. Especially compared to the chinese wires those develop green copper oxide just by sitting on the shelf, rendering them unsolderable...

One can also buy silver clad copper cables, those usualy come with PTFE jacket. It is a really cheap experiment to compare to similarly sized copper wires, whether skin effect makes different or not.

That won't make any difference with audio, as the copper underneath the silver plate will still carry most of the current, even at the high end of the audio band.

Just like Litz, I just had an idea for a cheap experiment if someone wants to try it.

I vaguely recall that I saw at some point a "speaker wire" made of flex PCB, where the two conductors were like 2" wide traces, while conductor quadrature were small but the surface was large.

Silver plated copper cable will not be cheap, especially if it's marketed for audio use.

Also how does one objectively test it? The listener needs to be unaware of which cable is connected when performing the test, otherwise cognitive bias will creep in: if they expect the silver to sound better for one reason, then they'll believe it.

[jmelson]

I vaguely recall that I saw at some point a "speaker wire" made of flex PCB, where the two conductors were like 2" wide traces, while conductor quadrature were small but the surface was large.

This stuff was mass produced for power connections to modular workspaces (cubicles) that ran under the carpet.  Nut sure it was 2" wide, but if it can handle 15A outlets, ought to be fine for audio.

Jon