Speaker Cable Faceoff 2: Introduction & Measurements

[Audioholics]

A year ago when we published our original Speaker Cable Faceoff 1 article, we had no idea that it would lead to almost consistent questioning of when we would post a follow-up with more cables and companies represented. Well, after months of careful measurement, analysis and observation, we unveil our second Speaker Cable Faceoff article. Using a Wayne Kerr 3260B Magnetics Analyzer we meticulously measured cables from several manufacturers and summarized our findings. If you thought our articles dispelling myths of dielectric absorption, cable distortion, cable resonance, and skin effect (in audible frequencies) were good - take a look at the measureable differences between cables from some well-known and some newcomer companies.

Read Speaker Cable Faceoff 2...

 

[Diablo]

Your article was very informative as far as it went. You left me hanging with no conclusions or rankings. Don't tease me, when do we see the rest?

 

[Unregistered]

Just read the article and have formed my own conclusion. The differences cited were teensy weensy and as described almost none of them would actually have any effect in the audible range.

So, other than appearance and quality of termination, why on earth would anyone pay $300/meter for a speaker cable when the AR zip cord measured practically the same and costs far less? The look of the axiom cable is nice and seems to be reasonably priced. I would pay a small premium for a speaker cable to get nice appearance and quality terminations, but no more than the cost of the axiom cables.

 

[zipper]

$1300 for a pair of 6 foot cables? I hope I don't look THAT stupid!

 

[Mudcat]

Your article was very informative as far as it went. You left me hanging with no conclusions or rankings. Don't tease me, when do we see the rest?

Sometime soon I hope, Gene has aske me to develop a ranking system based on weighted factors (resistance = 0.4; Inductance = 0.3; Capacitance = 0.2; and cost = 0.1), and a logirithmic formula. He did not give a time frame, and I was wondering when I will get the data or whether I'm to use the face-off article to cull the data?

 

[Audioholics]

The face-off articles are the data. Analysis is a separate article. It takes so incredibly long to write these articles and do the measurements that we deicded to break it up a bit.

 

[jneutron]

It takes so incredibly long to write these articles

wimps....

Cheers, John

 

[Dan Banquer]

Speaker Cable Faceoff

Hi John;
I count myself as one of those wimps, Nothing like a day job, part time job, and a family to cut into your time. However if anyone of you would like to fund a research project well enough so I can quit my day job and provide me with the necessary test equipment I will be happy to oblige.
I suppose the rest of you are now thinking to yourselves " when is this guy going to wake up from this dream?" or "what freakin planet did this clown fall off of anyway"
d.b.

 

[Unregistered]

Would be nice to have compared them with your previous winner, the Cobalt Cable. Although one of the cables i'm interested about was tested (Alpha Core), i'm disappointed that DH Labs, JPS Labs, BetterCables, CatCables, SlinkyLinks, Synergistic Research, Analysis Plus, & QED are not in the mix. Maybe in Part III?

 

[gene]

i'm disappointed that DH Labs, JPS Labs, BetterCables, CatCables, SlinkyLinks, Synergistic Research, Analysis Plus, & QED are not in the mix. Maybe in Part III?

Now that I have the new Wayne Kerr 6420 at my disposal 24/7 I will certainly be measuring more cables in the future. Bettercables refuses to send us anymore product since we singled out their bulletplug and poor termination connection as an example in our Component Video Cable article last year. Its too bad because simply fixing it would have been of great benefit to the enduser and little to no cost to anyone.

 

[Empirical Audio]

Issues and some interesting metrics

Gene and I have already had an email exchange on this, but I felt the need to clarify something. Namely that the inductance measurement that I made was off by 80% from Gene's measurement. The explanation is that the Ferrite terminators were not removed from the cable for this measurement. The cable itself has much lower inductance than Gene measured. Only the Goertz is lower than the Clarity7. Just as a Zoebel network should be removed before the measurements, the ferrites should have been removed as well.

Secondly, I propose a better metric for cable comparison: A product or weighted sum of products of the parameters.

Each of the parameters R, L and C were measured in the study, so each of them individually must have some level of importance. Also, the combined measurements have importance. For instance, extremely low inductance is not very interesting if the result is extremely high capacitance. Therefore, I propose the following figure of merit as a metric for comparison (50% weighting for R, 35% weighting for L and 15% weighting for C). If we compute the reactance for all three at both ends of the frequency spectrum:

HF Merit (lower is better) = R*(.50)+L*2*pi*20*(.35)+(1/(2*pi*20*C))*(.15)

= R*(.50)+L*(1.19)+(.0012)/C

LF merit (Lower is better) = R*(.50)+L*2*pi*20E3*(.35)+(1/(2*pi*20E3*C)*(.15)

= R(.50)+L*(43982)+(1.19E-6)/C

Overall merit = LF merit + HF merit

This covers both ends of the audio spectrum.

This weighting says that 50% of the performance of a cable is due to the resistance, 35% due to the inductive reactance and 15% due to the capacitive reactance. This weighting is just my best guess based upon my own experience.

It is fairly easy to crunch all of the numbers in the summary measurements to see which cables have the best overall merit. I will save this exercise for someone else, since I dont want to bias the result. Any takers?

 

[Mudcat]

Trust me, it's being done. But why limit it to just 20 & 20K Hz? The method I'm proposing averages all the measured frequencies between 20 & 20K Hz because we hope for a linear response within the range. So, since 100, 1K, 2.5K, 5K, 7.5K, 10K, 15K, & 20K were measured, I'm proposing to use the average. The averages will then be given a Ranking Factor between 10 (best) and 1 (worst) based on forward differences type of interpolation (I'm trying to set up a logitithmic interpolation), these ranking factors will be weighted based on an overall percentage which is very close to the percentage that you proposed to provide an overall weighted ranking.

Why pi?

And why in h&*# are you installing anything that would lower the performance - namely the ferrite terminators. Are these just added so you can jack up the cost to some obscene amount on the gullible masses? I mean $1765.00 for ten feet of wire that ain't as good a 10, 11, or maybe even 14 awg lamp cord. I'm in the wrong freakin business. Don't expect sympathy here.

When you use one word too much, it gets rank!

 

[jneutron]

Gene and I have already had an email exchange on this, but I felt the need to clarify something. Namely that the inductance measurement that I made was off by 80% from Gene's measurement. The explanation is that the Ferrite terminators were not removed from the cable for this measurement. The cable itself has much lower inductance than Gene measured. Only the Goertz is lower than the Clarity7. Just as a Zoebel network should be removed before the measurements, the ferrites should have been removed as well.

For the purposes of measuring the wire by itself, yes, you are correct. However, you are not selling just the wire, but a product..comprised of specific elements..the measurements as they stand give the numbers which will be seen in the end application.

It is good, however, to point out that the inductance is different as a result of the ferrite..as you consider the ferrite to be of use to the consumer..I do not argue the utility of it, just the confusion caused by spec'ing a cable in a way that is not consistent with what you are selling..

I think of it this way..my vehicle, and my friends, are both 200 HP... hers, a Sebring, mine a voyager..but what's attached changes the end app..I fly down the road like a pig, but I can carry a heck of a lot more...

Secondly, I propose a better metric.....

An interesting metric..and being tried..

I am working on a better metric, one that is based on energy storage and phase delays, but it is still quite a ways off...requiring quite a bit of test setup design.

RE: inductance/capacitance tradeoff..

The ideal transmission line will have L * C = 1031 * DC (relative dielectric coefficient). L in nanohenries per foot, C in pf per foot..the ideal will be coaxial, with inner conductor tubular.

That sets the minimum product that can be obtained, all designs based on a specific dielectric will follow that product, or if the geometry is inefficient, the product will be higher..never lower.

If you examine the graph, you can see the effect the geometry (or in your case, additional elements plus geometry) has on the product..for efficient geometries, these lines would be exactly the dielectric coefficients of the materials the wires are composed of.. As a result of inefficiencies and lumped elements, they are quite a bit higher in some cases, than the actual DC.

Your point that the elements skew the numbers is well taken...perhaps, in the future, both sets should be reported..as, higher inductance cable sets, by design, store more energy, and are more susceptible to external magnetic field loop pickup..while a lumped element does not necessarily do the same, and may indeed be a plus..

Cheers, John

 

[jneutron]

Why pi?

reactance is calculated using w

and w=2 * pi * f


Cheers, John
PS:sorry, couldn't get the html code to make a real equation..

 

[Mudcat]

reactance is calculated using w

and w=2 * pi * f

I always forget the basic stuff. Hey, I'm an ME. At least you haven't thrown the square root of -1 at me!

 

[gene]

Just as a Zoebel network should be removed before the measurements, the ferrites should have been removed as well.

Actually I did remove the Zobel for the Goertz otherwise the capacitance measurement would be out of whack and vary based on cable length.

I thought originally you told me those little beads were resistors, but I suppose even resistors have inductance so I do understand your concern. I didn't want to hack up your cable but I will update the article (as promised) with your commentary/concerns on this point.

This weighting says that 50% of the performance of a cable is due to the resistance, 35% due to the inductive reactance and 15% due to the capacitive reactance. This weighting is just my best guess based upon my own experience.

I agree with this approach and as Mudcat pointed out he has developed a weighing algorithm in Excel for me to accomplish this. Stay tuned for the analysis portion (don't hold your breath though) of my cable face off II article for the results.

And why in h&*# are you installing anything that would lower the performance - namely the ferrite terminators.

This is sometimes done at the output of an amplifier to help isolate the amp from reactive loading. Though I have heard of some deleterious effects when the L is high enough to interact with the speakers crossover and cause increased amplifier distortion. I personally believe if cable capacitance becomes excessively high, then Zobel termination would be a better route. But I suppose adding the ferrite could help reduce HF noise pickup in a hostile environment when plugged into an amp with a high unity gain crossing.

In any event, until John E proves me wrong (I am waiting John) I think its important to point out that even expensive cables which tout minimization of skin effect perform similarly to (audiophile frowned upon) 10AWG Zip cord.

 

[Empirical Audio]

"But why limit it to just 20 & 20K Hz? The method I'm proposing averages all the measured frequencies between 20 & 20K Hz because we hope for a linear response within the range."

Because you only have one set of measurements, and these were made steady-state at a particular frequency. Actually, only 20kHz matters. You can forget about 20 Hz.

"And why in h&*# are you installing anything that would lower the performance - namely the ferrite terminators."

Because the ferrites improve the performance in the typical system. If you read my theories on my webpage you will find that I believe, based upon experimentation, that transmission-line reflections in a low-loss cable can cause the impedance that the amplifier sees to fluctuate wildly, causing many amps to go non-linear over short periods, and thereby causing distortion. These are very special composition ferrites and they are fine-tuned to each cable. The cable without the ferrites is the one to measure because the ferrites do not simply add inductance to the cable. They behave as transmission-line terminators.

"I mean $1765.00 for ten feet of wire that ain't as good a 10, 11, or maybe even 14 awg lamp cord."

Seems to me that the measurements on the Clarity7 are significantly better than ZIP, and I guarantee you that the transient behavior is MUCH better that the average ZIP cord.

This is the area that is overlooked completely in this study. Steady-state measurements are insuffficient to describe the behavior of any cable or component. The dielectric absorption of the Zip cord is vastly different than that of my Clarity7. Once you have heard a high-performance cable like this in a resolving system, all other cables become inferior. Did anyone even bother to listen to any of these cables??????????

Steve Nugent
Empirical Audio
http://www.empiricalaudio.com

 

[Empirical Audio]

Gene wrote:
"I thought originally you told me those little beads were resistors, but I suppose even resistors have inductance so I do understand your concern. I didn't want to hack up your cable but I will update the article (as promised) with your commentary/concerns on this point."

Thanks, Gene, I would appreciate that. I never explained anything about the ferrites. It is really my fault for not thinking of this and providing a way for them to be removed.

BTW - These are not "filters" and they dont keep RF from getting into or out of the cable. They act specifically on the resonance of the cable. You can see many spectral plots of this on my website, with and without the ferrites.

 

[gene]

Steady-state measurements are insuffficient to describe the behavior of any cable or component. The dielectric absorption of the Zip cord is vastly different than that of my Clarity7

Steve Dielectric Absorption is another unfounded snake oil claim that has little relevence for speaker cables. Its so called effects are already accounted for with the Cp measurement.

See my brief article on this : Dielectric Absorption

As for transient response differences between your cable, vs Zip cord for example, they aren't much different and I suspect by simply inserting a Zobel at the end of the Zip Cord you can achieve similar response characterstics as I will demonstrate in my completion of the article.

Regarding amp stability and your inclusion of the ferrite, I do find merit in that, but god help anyone with an amp that is so susceptible to cable interaction. BTW, even with the inclusion of your ferrite, the inductance of your cable is still commendably low!

 

[Empirical Audio]

Gene - do you ever get to CES or THE Show in January in Las Vegas?

If so, you should come by my room - The St. Tropez #1203. I often do cable shootouts on-the-fly for people in the room. I bring 10 gauge Monster Zip cord and several medium-quality interconnects, such as Audioquest, Monster and Kimber for these shootouts. In every shootout so far, the people in the room were able to hear an improvement when using my cables. I would not even attempt this if I was not confident in them. I even challange anyone to bring their own cable for a shootout. Show me any other cable manufacturer that has the balls to do this at CES.