The reality of loudspeaker cable .... the one you cannot play with ....

[A.S.]

We noted recently that the signal from the microphone to the loudspeaker comprises a line. A line of electrons flowing towards the voice coil of the speaker drive units. Because the voice coils are in a magnetic field, and the voice coil is glued to a speaker cone, we can hear a sound.

Before we get too carried away with spending money on the speaker cable that connect the amp to the speakers, and which is sometimes the thickness of a hose pipe, what about the part of the chain that you the consumer can't play with? The voice coil wire itself.

Did you appreciate that the voice coil wire is not much thicker than a human hair? And if we unwind a voice coil from a RADIAL 5" or 8" woofer we can stretch it out a long way .... far longer than the cable typically used to connect your amp to your speakers! So, on the basis of the sound be limited by the weakest link - in this case the thinnest wire - I'm satisfied that QED79 strand (or equivalent) is 'more than adequate enough to get you going'.

[Labarum]

I love it!

But I hope you didn't destroy a perfectly good loudspeaker just the prove the point, Alan!

[jplaurel]

A couple of weeks ago, I would have taken Alan's post above for gospel. My Harbeths seem to sound just as good with all sorts of wire from my DIY quad conductor solid core stuff to the AV installer's standard 14 gauge Proflex and some 20 year old 16 gauge lamp cord I had laying around. But I'm starting to wonder if this "all wire is the same" business is true for all loudspeakers.

Some months ago, I set up a dual purpose home theater/listening room using a popular brand of hybrid electrostatic speakers. They worked very well for sound effects in movies, but for 2 channel music they were absolutely awful. The main gripe I had with them was in the midrange, which was thin, sharp and irritating. High notes from female vocalists could actually be painful to hear. I would turn the volume down to the point of it becoming background music - not at all engaging or involving. Most people pointed at the room acoustics. Fair enough, as I do have some work to do in that regard, but then why would my Harbeths and Magneplanars sound so good in the same positions in the same untreated room?

I could not bear to listen to the big electrostats any more and I was very close to putting them up for sale at a substantial loss. As reason failed, I turned to superstition and, in a last ditch effort, I decided to try some different wire. As luck would have it, I found someone in the local area selling some fancy wire with mystery boxes inline, and he was willing to let me try them out. I could hear Alan's posts about wire ringing in my ears as I went to a surreptitious meeting in a nearby parking lot.

I got the wires home and hooked them up. Lo and behold, there was a huge difference in the sound for the better. The midrange became enjoyable to hear, even on tracks with those penetrating vocals - no longer sharp and lean, but full, warm and more forward. The high end had more detail without being harsh. And there was more, tighter bass. I could not believe my ears.

My wife joined me and after listening to a few tracks of vocal music with eyes closed, she said "Ok, now let's hear the [electrostats]." Bingo! She was fooled into thinking she'd been listening to the Compact 7ES-3s, which I'd placed up front for comparison. When I told her she'd been listening to the ESLs, she was incredulous. The wire with the mystery boxes had made the speaker sound completely different in her mind. I know from reading the HUG that conclusions drawn from these sorts of informal comparisons are supposed to be all in our minds, that proper comparisons can't be made without real A/B switching with level matching, etc. But in my case, the difference is not at all subtle. With one set of wires, I was literally in pain, constantly fooling with the volume, eq, and just generally unhappy. With the new wires and their mysterious "network" boxes, the pain is gone, replaced with pure enjoyment.

So what gives? Are ESLs somehow more sensitive to wire types? Alan has said before that properly designed speakers are not sensitive to wire type. My Compact 7s certainly aren't. Maybe my ESLs are not well designed enough to be wire-independent? Or is is that the mystery box is just doing some EQ that I find appealing?

Mystified...

[Gan CK]

A couple of weeks ago, I would have taken Alan's post above for gospel..... I could not believe my ears. Are ESLs somehow more sensitive to wire types? Alan has said before that properly designed speakers are not sensitive to wire type. My Compact 7s certainly aren't. Maybe my ESLs are not well designed enough to be wire-independent? Or is is that the mystery box is just doing some EQ that I find appealing? Mystified...

Hi Jplaurel, yes your experience is mystifying indeed! Some years back a friend of mine wanted to sell me his very expensive pure silver interconnects & he was very confident that it would beat the pants off my cheap Qed interconnect. So he brought the pure silver I/C to my place & after several repeated comparisons, 3 pairs of ears there couldn't discern any difference at all. None at all. But when i was fiddling with another pair of interconnect on another day, i could hear a subtle difference. But the difference was so subtle that it didn't really seem to matter at all.

[A.S.]

...So what gives? Are ESLs somehow more sensitive to wire types? Alan has said before that properly designed speakers are not sensitive to wire type. My Compact 7s certainly aren't. Maybe my ESLs are not well designed enough to be wire-independent? Or is is that the mystery box is just doing some EQ that I find appealing? Mystified...

With respect, there are three issues in your observation which, IMHO pose some very serious questions. When presenting such exciting opinions under the Harbeth umberella of credibility, we ask that proper heed has been take on our methodology stated here.

1) What on earth is the 'mystery box'? Unless you understand what this is doing, how it alters the electrical signal - and why - then you are not commenting on the audio characteristics of speaker cable in isolation but speaker cable plus 'mystery box' in combination. You need to remove or bypass that box entirely to be able to comment on speaker cable alone. Do you see the importance of doing that? My guess is that the box contains some load-compensating circuitry.

2) By bringing electrostatic speakers into the discussion (and we are not able to really discuss electrostats as we don't make them) you have introduced a class of speaker which presents a notoriously difficult and complex electrical load to the amp. This is because you can not drive the electrostat's diaphragms directly from the amplifier as you can with a conventional moving-coil speaker. You have to drive the panels through (usually) a large, inductive, load variable matching transformer. And that transformer presents a weird load to the stereo amp. It is therefore important to appreciate that the electrostat panel + its internal matching transformer or other circuit + the speaker cable + the amp will form a complex interdependent system. Change any one part and I would fully expect there to be a change in sound. Conventional speakers have a relatively benign impedance of perhaps around 6 ohms: electrostatics can have a load as low as 0.5 ohm (ten times lower) and/or highly reactive and/or peaking at some very high value.

3) What would complete your observation in a way worthy of fuelling excitement here would be to try those wonder-cables on the C7s with/without the magic box and/or constructing the entire tests in the way recommended above. Can you see that you really should have attempted that before exciting readers here?

I have no idea of which Magneplanar speaker you have, but just to give you the idea of how complex the speaker load presented to the amp can be, here is a caution taken at random from their web site ...

"Check the specifications for your amplifier, or one you may be considering, to determine if it is rated (watts) for 4Ω speakers. Virtually all amplifiers are rated for 8Ω speakers. Many are also rated (watts) for 4Ω with a few rated for 2Ω and 1Ω. If you are purchasing a new amplifier to use with the MMGs, you should purchase one rated for a minimum of 4Ω. If you own a receiver rated into a 8Ω load, and there is no mention of a 4Ω rating...BE CAREFUL! Driving some of these amps hard can cause damage to them if they are not well-fused. If your amplifier manual makes no mention of a 4Ω load, check with the manufacturer. Magnepan is not responsible for damage to amplifiers."

This emphasises the unusual load 'stats typically present to the driving amplifier. They are not conventional speakers and they do not present a conventional load! They are, in essence, gigantic capacitors with plates that vibrate and produce sound! Their electrical load, and maybe their sound, is likely to be amplifier and cable dependent!

And to look at electrostatic speakers and their load: I have to hand a link to a review of the QUAD electrostatic here. From the bottom plot of resistive impedance (phase not shown), you can see that the curve looks very different to that of a conventional speaker in overall shape. But stressing my point about the complexity of load presented by electrostatic speakers, note how unlike a conventional speaker, the impedance curve varies depending upon the audio signal voltage applied to the speaker's input terminals. And inside a 'stat .... complex transformers and electronics (pix).

Unwittingly "The Myth" has been given another puff of oxygen.

-----------------------------------------------------------------

P.S. If you notice at the foot of the page it says 'we do not encourage discussion of cables etc.'. I'm going to discuss with Moderation future handling of 'cable posts'. Even discussing cables can unwittingly endorse them, when tests we ourselves have conducted cannot when driving load-benign Harbeth speakers. We reiterate our corporate position on cables etc. as stated here.

[KT88]

Just to clarify, Magneplanars are not electrostatic speakers, and in his post, jplaurel mentions that the Harbeths and Magneplanars both sounded good. He mentions hybrid elctrostatics, so I'm guessing Martin Logan?

[HUG-1]

Thanks for the correction. I'll rework my post to minimise confusion. (Done, I hope clear now)

The point I was attempting to make, and in this regard it doesn't matter whether the speakers were ribbons, pure electrostatics or hybrid electrostatics, is that the load presented by different speaker technologies must be considered. And that load cannot be deemed relatively or very benign as it is with conventional moving coil speakers. The further the load is from behaving like a standard resistor the more interaction wthere will be with the amp, and logically, the speaker cable will form part of the driving circuit. Speakers employing some sort of audio-input-to-panel-element-matching (a transformer) like the QUAD 'stat, and I assume most/all other electrostats, will present the amp with some sort of odd-ball load. And that load can't be assumed to be trivial. And hence, strange things could well happen depending upon the exact amp/speaker cable/speaker. And yes, if the speaker has a difficult load and is driven from an amplifier that is load-sensitive the speaker could sound different depending on the amp/speaker interface. No doubt about that (in theory).

Be that as it may, the point we here at the official Harbeth User Group are solely concerned with is how does the speaker cable effect the sound of the Harbeth speaker - if at all. Since we have said, innumerable times, that 'you can use any Harbeth with any competent amp' (i.e. all fully-functional hifi amps) it follows that the speaker load must be and is benign, and it folows then that the potential for the cable to play a significant part in shaping the overall sound, negligible to zero. And armed with what we consider to be AB tests that prove that to our satisfatction (or did, last time we looked at it), is our position.

(Alan is deputy moderator today)

[jplaurel]

I want to clarify that it wasn't my intent to endorse exotic wires or perpetuate the myth surrounding them. Some quick answers to Alan's questions:

1) A little digging reveals that the inline box on the cable is something called a "Zobel network". From what I gather, these were used to EQ telephone lines. This link has a photo captioned "BBC engineers equalising audio landlines circa 1959. The boxes with two large black dials towards the top of the equipment racks are adjustable Zobel equalisers."
http://en.wikipedia.org/wiki/Zobel_network

2) Electrostats are indeed way off topic here. The specs for mine indicate they can dip as low as .8 ohms.

3) I have indeed tried all the wires at hand on my C7s and am quite certain that I could not hear any difference between them in a blind test. Indeed, the reason for my post was to understand why this was the case with the C7s, but not with the ESLs. Alan's comments about ESLs were very helpful in this regard.

......1) What on earth is the 'mystery box'? Unless you understand what this is doing, how it alters the electrical signal - and why - then you are not commenting on the audio characteristics of speaker cable in isolation but speaker cable plus 'mystery box' in combination. You need to remove or bypass that box entirely to be able to comment on speaker cable alone. Do you see the importance of doing that? My guess is that the box contains some load-compensating circuitry.

2) By bringing electrostatic speakers into the discussion (and we are not able to really discuss electrostats as we don't make them) you have introduced a class of speaker which presents a notoriously difficult and complex electrical load to the amp. This is because you can not drive the electrostat's diaphragms directly from the amplifier as you can with a conventional moving-coil speaker. You have to drive the panels through (usually) a large, inductive, load variable matching transformer. And that transformer presents a weird load to the stereo amp. It is therefore important to appreciate that the electrostat panel + its internal matching transformer or other circuit + the speaker cable + the amp will form a complex interdependent system. Change any one part and I would fully expect there to be a change in sound. Conventional speakers have a relatively benign impedance of perhaps around 6 ohms: electrostatics can have a load as low as 0.5 ohm (ten times lo...etc....

[A.S.]

1) A little digging reveals that the inline box on the cable is something called a "Zobel network". From what I gather, these were used to EQ telephone lines. This link has a photo captioned "BBC engineers equalising audio landlines circa 1959. The boxes with two large black dials towards the top of the equipment racks are adjustable Zobel equalisers."
http://en.wikipedia.org/wiki/Zobel_network......3) I have indeed tried all the wires at hand on my C7s and am quite certain that I could not hear any difference between them in a blind test. Indeed, the reason for my post was to understand why this was the case with the C7s, but not with the ESLs. Alan's comments about ESLs were very helpful in this regard.

Ok, noted.

Now there are a couple of follow-up points.

First, the 'Zobel network' is nothing more fancy than a capacitor, resistor and maybe a coil in circuit. Second, as you say, the Zobel circuit (named after a Russian I think) was indeed used to 'equalise long phone lines', typically in the hundreds of miles range. But what's the relevance of that? Electrically a long phone line is a completely different animal to a short speaker cable. For a start, phone lines have a 600 ohm impedance and speaker cables almost zero impedance. So any Zobel-like techniques designed to work on a 600 ohm miles-long telephone wire will not have the intended result at all without much adaptation and a total redesign with a speaker/cable/amplifier load. The values of the components in the a Zobel network designed for a telephone line will be in error by factors of hundreds or thousands or more when used in an amp/cable/speaker arrangement. I'd expect that your Zobel box is optimised for particular speaker/(amp)/cable set-up but it probably wouldn't be suitable for any other combo.

You'll note that I said:

My guess is that the box contains some load-compensating circuitry.

And I was right. That's other words for 'Zobel network'.

So what about the Zobel network. What does it actually do? Well, believe it or not, there are Zobel networks in Harbeth speakers. Yes - right: this is not magical technology at all. We use it. And we've written about it here in tech Talk. I've even given you the equations to calculate appropropriate values. The point is this: all the Zobel network does if (if implemented correctly) is to partially neutralise the oddities of the impedance of the speaker (+ cable) as seen from the amp looking out from its output terminals. It makes that combined load look more like a resistor - less reactive and easier to drive.

Questions I'd have asked if I'd been with you making this experiment:

1) Is the improved sound you've with certainty attributed to the wonder-cable achievable without the magic box, Zobel equaliser? Very unlikely.

2) Would that same Zobel box when used with another (standard) cable also give a better overall sound? Probably yes for the same speaker providing that the Zobel box is designed to neutralise the speaker alone and not neutralising the speaker + some weirdness of the new fancy cable. If you change the speaker you must redesign the Zobel circuit completely. If there is something electrically unusual about the speaker cable this too may be compensated-for in the Zobel. We can't be sure.

3) Would that Zobel + cable improve the sound of the C7? Exceedingly unlikely. The Zobel + cable would need to be designed to neutralise a specific load, a specific model of a specific speaker using a specific length of cable, and electrostats have a completely different load to conventional speakers. And anyway, you don't need to equalise the benign load of a Harbeth because I've taken care of that in the design for your amp's benefit.

4) What do your observations highlight? Have we really disproved what A.S said about cables? No you have only emphasised the point that certain non-Harbeth speakers can and do have complex loads and these loads can and will have an interaction with the driving amp and the load can, with care, be ameliorated and this will/may/might have a sonic effect before/after load neutralisation. This is science.

5) What is the implication for the amplifier/speaker interface? If the speaker (and/or cable) presents a difficult load to the amp, the voltage appearing at the speaker's terminals may be unpredictable and unstable and certain sounds would be boosted, others suppressed. The speaker could ring at certain frequencies if the amp cannot drive the strange load.

6) Does any of this prove that you have heard a difference in sound due to the cable alone? There is not one scintilla of evidence from the test as conducted that what you have heard - beneficial or otherwise - can be attributed to the cable as you have not isolated the influence of the cable from that of the Zobel box. In other words, you have introduced two variables into your experiment and cannot draw any safe conclusions because the experiment lacks intellectual rigour. See here (again).

Conclusion: you cannot draw valid observations that are fit to influence others from what you have heard and thus your comment:

A couple of weeks ago, I would have taken Alan's post above for gospel...

should not be challenged on the basis of your experiment, as described. More careful experimenting may indeed disprove your previously held opinion (i.e. mine) but not this one.

[A.S.]

P.S. One thing to watch with parallel Zobel networks: although they flatten the load, the resulting load after the Zobel is applied, is always lower than the lowest impedance pre-Zobel.

For example, if the speaker load varied from 87 ohms maximum at some frequency to to 3.2 ohms minimum at another, after the Zobel is fitted the 87 ohm may be reduced to under 10 ohms (a guesstimate) but the 3.2 ohms may now dip to 2.5 ohms. So the Zobel circuit is useful for clobbering a rising impedance (which most amplifiers are highly tolerant of anyway) and actually makes the impdedance minima (which amps don't like) even more severe. It is not a universal solution and rarely if ever used in the bass/midrange band where much power is drawn from the amp! Again, if a speaker presents a 0.8 ohm load at some frequency, with the Zobel this could drop to an eye watering 0.3 ohms or lower ... and that presents the amp with a virtual short circuit. No wonder the amp/cable/Zobel/speaker interface can have an influence with difficult and/or very low impdedance loads. BUT NOT WITH HARBETH SPEAKERS.

[jplaurel]

If there's one thing I've learned from this thread, it's that I need to communicate with more precision. I've given the impression of disagreement where none exists. To Alan's points:

1) Agree that it is very unlikely there would be any difference in the sound with the wire sans the Zobel equalizer. The difference I'm hearing is most certainly due to the Zobel.

2) Once again, I agree that it is the Zobel making the audible difference. From my limited understanding of the Zobel concept (I am a software engineer with no EE training), I understand that this type of circuit must be designed/calibrated to take into account the amp, speakers, cable length and properties of the wire itself. In this case, I probably just had the dumb luck of stumbling onto a cable+Zobel combo that just happened to work well with this particular speaker and amp.

3) I have, in fact, tried this cable + Zobel with my Harbeth C7s and cannot discern any audible difference between it, plain 16 gauge zip cord or any other wire for that matter. My C7s are connected with standard, commercial-grade 14 gauge stranded copper wire. They sound equally wonderful driven by expensive tube and SS amps or the simple 30W Peachtree integrated dac/pre/amp/ipod dock that I use on my desktop.

4) My original post was not meant to challenge what Alan has been saying about cables with Harbeth and other properly-designed speakers. However, I was questioning whether the same is true for electrostatics. I was very interested to learn that some non-Harbeth speakers and ESLs in particular can present very strange, complex loads which can affect sonics.

5) Alan's comment that unpredictable voltage appearing at the speaker terminals could cause the speaker to "Ring at certain frequencies" is very interesting. That's a good description of the irritating sound I was hearing from the ESLs. I assume the Zobel is at work here, flattening the load and thus reducing this ringing effect.

6) Agree completely. I have not provided any evidence that the cable alone is making the difference and apologize if I inadvertently made that implication. And just so everyone is clear, I'll say it once more: I agree that any difference I'm hearing is certainly due to the Zobel. Unfortunately, I can't test the cable & Zobel separately without butchering the cable assembly.

Alan: Thank you very much for taking the time, which you can scarcely afford to spend on a topic like this, to offer your comments.

Now let's go listen to some music!

Questions I'd have asked if I'd been with you making this experiment:

1) Is the improved sound you've with certainty attributed to the wonder-cable achievable without the magic box, Zobel equaliser? Very unlikely.

2) Would that same Zobel box when used with another (standard) cable also give a better overall sound? Probably yes for the same speaker providing that the Zobel box is designed to neutralise the speaker alone and not neutralising the speaker + some weirdness of the new fancy cable. If you change the speaker you must redesign the Zobel circuit completely. If there is something electrically unusual about the speaker cable this too may be compensated-for in the Zobel. We can't be sure.

3) Would that Zobel + cable improve the sound of the C7? Exceedingly unlikely. The Zobel + cable would need to be designed to neutralise a specific load, a specific model of a specific speaker using a specific length of cable, and electrostats have a completely different load to conventional speakers. And anyway, you don't need to equalise the benign load of a Harbeth because I've taken care of that in the design for your amp's benefit.

4) What do your observations highlight? Have we really disproved what A.S said about cables? No you have only emphasised the point that certain non-Harbeth speakers can and do have complex loads and these loads can and will have an interaction with the driving amp and the load can, with care, be ameliorated and this will/may/might have a sonic effect before/after load neutralisation. This is science.

5) What is the implication for the amplifier/speaker interface? If the speaker (and/or cable) presents a difficult load to the amp, the voltage appearing at the speaker's terminals may be unpredictable and unstable and certain sounds would be boosted, others suppressed. The speaker could ring at certain frequencies if the amp cannot drive the strange load.

6) Does any of this prove that you have heard a difference in sound due to the cable alone? There is not one scintilla of evidence from the test as conducted that what you have heard - beneficial or otherwise - can be attributed to the cable as you have not isolated the influence of the cable from that of the Zobel box. In other words, you have introduced two variables into your experiment and cannot draw any safe conclusions because the experiment lacks intellectual rigour. See here (again).

Conclusion: you cannot draw valid observations that are fit to influence others from what you have heard and thus your comment:

should not be challenged on the basis of your experiment, as described. More careful experimenting may indeed disprove your previously held opinion (i.e. mine) but not this one.