Speaker Cable Faceoff 2: Introduction & Measurements

[jneutron]

Hey, I'm an ME.

Oh...sorry to hear that...I'll ratchet it down a bit...
(oh man, when are you guys gonna stop tossing those straight lines at me??)

In any event, until John E proves me wrong (I am waiting John)

Let's see...what was it Hawk said??? OH yes...

It takes so incredibly long to write these articles and do the measurements that we deicded to break it up a bit.

At least I have an excuse...I just mixed and poured 8 * 60 lb bags of sackrete in my basement lab. getin it ready to do anything at all..next is walls and workbench...think I'm gonna use some butcher block pine pre-made from da depot.. (the excuse is not because I don't have time, but that I'm in a back brace from the quarter ton of ceeeement I carried into the basement.

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.

Yes, indeed..they certainly do..they absolutely provide excellent amplitude and voltage phase delivery to the load from 20 to 20K...what I am questioning, which I believe is consistent with Steve's belief (although I do not know if he understands what I have been speaking of..) is that the phase delays from the different cables will affect the soundstage...

So my premise is that normal 20 to 20K testing, of any sort, does not see what I am looking for...

That was why I complained to you about the first runoff, where you didn't go high enough in freq....but now, you going to 1Mhz? I'm a happy camper...

Hey..I use #12/3 extension cord...rugged, colorful, can drive a forklift over them without failure...meets the need..mobile app, of course...

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.

Yah, yah...I've heard that one before... (probably because I've been saying almost the same thing...)..

Hi Steve..

I have found no real studies that test specifically for human lateralization with two sources, time shift vs directionality in particular...I do find the latest multispeaker array/multiprocessing stuff of interest, but I fear the study results will remain captive in house until the market opens up for the new technology.. (forcing me to setup my own tests)

I find no studies of lateralization cue shifting as a result of cable parameters.. that would tie what you (I) am saying...to what you are hearing..

From a scientific point of view, you have not shown your product to be of any advantage over zip...you have anecdotal accounts...but nothing else..

So, quite honestly...Gene's article is applying what is currently known to cable measurements....I certainly can't fault him for that...but, that doesn't mean I agree that his tests are sufficient...You are incorrect in your statement that steady state measurements cannot be used...they most certainly can, but there needs to be a model tie in between that and what is perceived..that conceptual gap is a hole large enough for a tractor trailer..and I am the NUT holdin onto the steering wheel...

Guys, I'll say this once....

ALL THE TESTING YOU GUYS ARE DOING IS MONOPHONIC BASED TESTING......UNO...ONE...SINGLE...
I DO NOT LISTEN WITH ONLY ONE EAR...NOR TO ONLY ONE SPEAKER..

Think about what is entailed when two channels carry the information, and they have to be timewise accurate to within about 2 to 5 uSec...

Mono sources do not care a whit if some of it's content is timeshifted 5 uSec...but we can here that if it's between ears..

..

That is where it's at...I'll get there...perhaps a slowboat to china, but I'll get there...

Cheers, John

 

[Mudcat]

Mono Signals vs Stereo

John,
Remember my post a couple of months ago about using 12/4 cable to carry both channels to a subwoofer?

I will say that I did make some tests using both a pink noise signal, and a high power test with a 500 watt photo lamp. However (and Gene may freak at this), I have to invalidate all those measurements (yeah, even the stuff I sent you Gene on those DIY cables). I just found out - 1 hour ago, honestly - my LCR meter requires recalibration (I've got to stop using expensive electronics as a hammer). Will report and rewrite when able.

Now I'm getting a seriously good edumacation here, so what controls/affects phase delays? Reactance?

 

[gene]

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.

Steve what type of "resonance" are you referring to? Mechanical or electrical? I am hoping you are referring to mechanical since cables don't electrically resonate. Stereophile/Trasparant Cable unfortunately promotes that fallacy, but it just isn't true.

Speaker Cable Resonance

As for your DBT tests at CES, I commend your efforts and yes we do goto CES, see our CES 2004 Coverage We will be at Cedia next month, will you be there?

BTW, I never claimed cables do not sound different. However I will argue that only poorly designed cables are sonically different and definately measurable.

That was why I complained to you about the first runoff, where you didn't go high enough in freq....but now, you going to 1Mhz? I'm a happy camper...

John I don't get your beef? Even Speaker Cable Face Off I tested out to MHz. I always test out to at least 1MHz.

At least I have an excuse...I just mixed and poured 8 * 60 lb bags of sackrete in my basement lab.

Let me play my violin for you John I spent last weekend building an extension on my home while working on this site till 2am each night only to get up the next morning at 7am to goto my Cube and then start the whole process over again. Weep weep

 

[jneutron]

John I don't get your beef? Even Speaker Cable Face Off I tested out to MHz. I always test out to at least 1MHz.

Hmmmm...I remember some data that didn't go out there...of course, now, I can't find it...didn't your first e-mails to me stop befor a meg?

Let me play my violin for you John I spent last weekend building an extension on my home while working on this site till 2am each night only to get up the next morning at 7am to goto my Cube and then start the whole process over again. Weep weep

Wimp..

Man, I do know that process, though..did that on my house bout 7 years ago..gonna be doin it again next year...

But, Gene...I gotta tell ya...you rule work, not the other way around..don't get too far in it that it kills ya..

I have to invalidate all those measurements (yeah, even the stuff I sent you Gene on those DIY cables). I just found out - 1 hour ago, honestly - my LCR meter requires recalibration .

Past cal date? Or wildly off?..you should have a set of standards available to check the accuracy of your meter..

In fact, that is actually a really good idea..we should take some wire, cut ten foot lengths, measure accurately (gene, ya listenin?), and distribute them as standards...and occasionally, have Gene re-verify them...will save the heartache you are now experiencing..

Now I'm getting a seriously good edumacation here, so what controls/affects phase delays? Reactance?

Yes. it is consistent with that of a spring suspended mass..pull up on the spring, the mass has a delayed reaction...consistent with inductance..introduce a shock absorber...capacitor..spring constant...resistor..all interplays to define what the instantaneous position of the mass is going to be.

An inductor can have an instantaneous voltage step across it, say from amp to speaker..the inductor will charge, and current will increase, based on the inductance and the voltage...dI/dt=V/L....meaning, the larger the inductance, the smaller the rate of change of current..that slowing of the signal rate of rise...the distributed capacitance also affects, but now the relation is dV/dt=I/C..meaning, the cable capacitance will shunt current as a result of fast voltage changes..the net result, for example, on a square wave, would be slowing down the rise and fall speeds..

I worry about that w/r to lateral imaging..

Cheers, John

 

[Unregistered]

Gene wrote:
"Steve what type of "resonance" are you referring to? Mechanical or electrical? I am hoping you are referring to mechanical since cables don't electrically resonate. Stereophile/Trasparant Cable unfortunately promotes that fallacy, but it just isn't true"

Really? Then how do all of the amplitude peaks appear in the spectral plots that I have done for my cables and others? Please look at my FAQ page under resonance to see these plots:
http://www.empiricalaudio.com

 

[Empirical Audio]

Gene wrote:
"As for your DBT tests at CES, I commend your efforts and yes we do goto CES, see our CES 2004 Coverage We will be at Cedia next month, will you be there?"

No, we only go to THE Show and occasionally to the Stereophile show, if it is not in NYC. Even CES is a bit of a waste for us, except for getting the reviews. We dont sell much through dealers and dont want to. Mostly have had unpleasant experiences with them. Mostly just a small business that is dominated by component mods currently and we sell a few cables direct.

 

[jneutron]

Gene wrote:
"Steve what type of "resonance" are you referring to? Mechanical or electrical? I am hoping you are referring to mechanical since cables don't electrically resonate. Stereophile/Trasparant Cable unfortunately promotes that fallacy, but it just isn't true"

Really? Then how do all of the amplitude peaks appear in the spectral plots that I have done for my cables and others? Please look at my FAQ page under resonance to see these plots:
http://www.empiricalaudio.com

Gene is obviously refering to resonance in the audio band..Steve, cable length resonance, 9mHz in the FAQ section.

Steve: have you any real data showing the phase of the current at the load with the speaker attached.

If so, do you have data for varying cable impedance vs phase plots of current at the load..

Are you interested in doing some collaboration?

"CES" ...if it is not in NYC

bummer..t'would be a nice discussion..you guys over there...geeze, don't you know, the east coast is the center of the universe??

Cheers, John

 

[Mudcat]

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

I will save this exercise for someone else, since I dont want to bias the result. Any takers?

You are biasing the results by proposing this format since Resistance and Capacitance will have a negligible impact. At 20 KHz, a high capacitance cable, of which your is one (not the highest, but high none the less), will have a reactance of something like 10^-8 Ohms, and 10^-5 at 100 Hz. Your cable would climb from 6^th place to maybe third in my ranking system, which generates a non-dimensional factor for all the parameters measured.

In the final analysis conducted by someone out to buy a speaker cable, they will have to weigh the benefits of spending $1765 on 10 feet of wire that came in 6^th place (by my ranking method) vs less than $100/10ft for something that performed better.

so, if you really want to reduce capacitance to a non factor, may I suggest some AR or Sound King zip wire.

And if you really want to compete with Inductance, Try one of my DIY Cat 5 cables.

 

[gene]

Steve;

Really? Then how do all of the amplitude peaks appear in the spectral plots that I have done for my cables and others? Please look at my FAQ page under resonance to see these plots:

Cables do NOT electrically resonant! They are not like a basic RLC lumped circuit in that respect.

Please take a look at Rod Elliots article that we worked on together in the past titled Loudspeaker Cable Characteristic Impedance.

What you are showing is interaction between the amplfier and complex impedance of the speaker cable + loudspeaker impedance. A simple zobel (50cents) attached at the end of the speaker cable to match the charactersitic impedance of the cable will resolve this nicely. It is likely not a real world problem (unless the amplifier runs out of phase margin) especially since any decent amp is bandwidth limited below unity gain crossing at 10MHz, not to mention brickwall filtering of your digital playback device. Most speakers do have integrated zobels to control high frequency impedance to limit any amplifier energy that may occur from this type of resonance interaction.

Mud;

And if you really want to compete with Inductance, Try one of my DIY Cat 5 cables.

I am looking forward to publishing your articles when you complete the measurements. Those cables certainly are a great value and measure very well! No zobles or ferrites needed either

 

[jneutron]

Steve;
Cables do NOT electrically resonant! They are not like a basic RLC lumped circuit in that respect.

Um, Gene? he is talking about the resonance that occurs when the terminations are not equivalent to the characteristic impedance of the line..in which case, you have to deal with the reflection coefficients..this is NOT a transverse resonance, it is an axial one based on prop speed of the effective dielectric.

He is measuring that, and his faq dialogue on resonance specifically showed that. In addition, he overlayed two plots, one with clarity 7 cable, red is 3 ohm termination, green is 11 ohm..(the correct termination impedance), which completely removed any reflection effects up to about 25+ Mhz..

Steve: it would have been nice to have provided figures on your faq section...instead of me saying "you know, that picture"...

What you are showing is interaction between the amplfier and complex impedance of the speaker cable + loudspeaker impedance.

he is demonstrating the time of flight characteristics of a 4.5 meter cable that is terminated with mismatches..the positive reflection coeff when the load impedance is > than the line, and the negative one when it is lower. The coeffiecients at each end will determine the Q of the length of wire as an electrical resonator.

A simple zobel (50cents) attached at the end of the speaker cable to match the charactersitic impedance of the cable will resolve this nicely.

I think steve is already doing that..with the ferrite thingy.

Mudcat: it will be interesting to see the L/C tradeoff you are building into the cat 5 braid...attached is a graph which shows the absolute best LC tradeoff that can be realized with coaxial structures, which btw, are the most efficient method of power transfer of electricity in conductors..other than superconductors..I am building a big graph which I will plot all cables onto, showing relative position on the chart...

As to the rating system being chosen: What are the weights, and what justification do you have for choosing them specifically..

Cheers, John

 

[Mudcat]

The Cat 5 wires literally suck hind*%$ when it comes to capacitance. And to be fair to Steve, they are probably no better in inductance either. I was just on a rant after figuring out his proposed comparison artifically improve the Clarity position compared to others.

 

[jneutron]

The Cat 5 wires literally suck hind*%$ when it comes to capacitance. And to be fair to Steve, they are probably no better in inductance either. I was just on a rant after figuring out his proposed comparison artifically improve the Clarity position compared to others.

Ummm, mudcat....what does suck hind*%$ mean? High capacitance?

His proposed comparison, if I have read correctly, reflects what he has stated is his experience in what matters for him when he builds his wires and tests them...whether it is skewed to favor him or not is not my concern...it begs the same question I posed to you...what justification is given to the weights used..

I can build a cable that exceeds every single one I've seen so far, and can build them to far surpass his, yours, JR's...everyone...it doesn't mean that what I build is even worth a damn....just that I have the ability to dial the inductance to less than 250 pHenries per foot, and can keep the capacitance the lowest, while still making the cable outta #4 welding wire guage...

But, how does the merit number reflect the application? Don't know..

I personally would use a merit equation that includes power dissipated as well as power detained.

Cheers, John

 

[Steve Nugent]

You are biasing the results by proposing this format since Resistance and Capacitance will have a negligible impact. so, if you really want to reduce capacitance to a non factor, may I suggest some AR or Sound King zip wire.

And if you really want to compete with Inductance, Try one of my DIY Cat 5 cables.

Resistance and Capacitance - a negligible impact? I'm afraid I'll have to disagree on this.

And I have tried lots of ZIP and CAT5 cables etc... years ago. They suck compared to my current designs. I believe some of you guys need to upgrade your systems so you can hear some of waht I am hearing. Many audiophiles feel that they have reached audio nirvana and cannot go any further, only to find that with mods, they can go several levels higher in performance. Unless you have REAL DEEP pockets, you cannot get there without mods IMO.

 

[gene]

John;

Again Steve's FAQ clearly states that the cable is resonanting as if it were a lumped RLC circuit and proposes a solution "

Empirical Audio has taken steps to deal with resonance in our cables.

Cables don't resonant. Again what we are seeing here is the amplifier + cable + speaker interaction at frequencies where the characteristic impedance mismatch between the TL (Cable) is allowing for reflection.

The dielectric material does in fact influence just what length that will occur due to propogation velocity as you inferred.

Simply terminating Zip Cord with the proper zoble will nullify this simulated, but highly unlikely real world problem.

 

[Unregistered]

Steve: have you any real data showing the phase of the current at the load with the speaker attached?

If so, do you have data for varying cable impedance vs phase plots of current at the load..

Are you interested in doing some collaboration?

Real data? No. Only computer simulation data on this.

Collaboration? I wish I had the time. Between building and selling cables and doing mods, I am busy 24/7.

 

[Unregistered]

I was just on a rant after figuring out his proposed comparison artifically improve the Clarity position compared to others.

Really? I had not done the calculation yet. It does not surprise me. The concept with Clarity7 is that you get inductance almost as low as Goertz without paying the high-capcitance penalty.

 

[Unregistered]

Cables don't resonate. Again what we are seeing here is the amplifier + cable + speaker interaction at frequencies where the characteristic impedance mismatch between the TL (Cable) is allowing for reflection.

John - if cables dont resonate, then why is the value of my ferrite a function of the length of each cable????

 

[jneutron]

John - if cables dont resonate, then why is the value of my ferrite a function of the length of each cable????

Ummm, Steve? I think you meant Gene here..

The concept with Clarity7 is that you get inductance almost as low as Goertz without paying the high-capcitance penalty.

hmmm..the first cable I made had 9.7 nH per foot, and 288 pf per foot...that was the design target, and what was measured..if I so choose to compromise inductance to achieve the goertz 20 nH number, I'll have 140 pf per foot instead of their 327 pf..or, if I build to your 34 nH number, I'll have 82 pf per foot..where you have 118 pf..Or, conversly, if I choose 118pf, I will achieve 24 nH per foot..vs your 34..btw...how do you test that low an inductance? It's always a bear..

Cables don't resonant. Again what we are seeing here is the amplifier + cable + speaker interaction at frequencies where the characteristic impedance mismatch between the TL (Cable) is allowing for reflection.

Gene...resonance is the storage of oscillating energy..in the transmission line world of engineering, they design and use lengths of transmission line to form resonating structures..the 1/4 wavelength transformer is a classic example, where they match two different impedances with a 3rd one...the actual quarter wavelength piece is working in resonance to adjust the I/V characteristics of the other two transmission lines.. so it is, in fact, a resonance...

The dielectric material does in fact influence just what length that will occur due to propogation velocity as you inferred.

Ummm...actually, I said ""based on prop speed of the effective dielectric.""..so I did not infer it, I stated the relation directly..


Cheers, John..

PS..hows the house coming?

 

[gene]

Gene...resonance is the storage of oscillating energy..in the transmission line world of engineering, they design and use lengths of transmission line to form resonating structures..the 1/4 wavelength transformer is a classic example, where they match two different impedances with a 3rd one...the actual quarter wavelength piece is working in resonance to adjust the I/V characteristics of the other two transmission lines.. so it is, in fact, a resonance...

Yes I am well aware of 1/4 wavelength distributed transformers though I don't consider that to be the same as electrical resonance exhibited in RLC networks.

The problem is evoking a lumped circuit model for a distributed network (cable). When you calculate cable resonance using a single RLC network for a cable it is based on the total inductance, capacitance and resistance of the cable. This is not how it works. Do the math and you will find as you break the cable into infinite RLC networks, the resonance frequency becomes infinity.

In actuality simulating a cable as a single RLC does the cable a disservice since you achieve minimal accuracy. The question is, to what precision would you like to model the cable? IE( .1dB, .001dB?) Simply start with a single RLC representation of the cable and model its losses into a known load, then break the cable into multiple RLC’s until you achieve the limit of the accuracy you are looking for.

PS..hows the house coming?

I haven't worked on the house this week since I am working here. I am hoping to paint next week and close in the room the following week so I can move all my test gear in there
(if the wife lets me)

 

[jneutron]

Yes I am well aware of 1/4 wavelength distributed transformers though I don't consider that to be the same as electrical resonance exhibited in RLC networks.

Ah, ok...there is the issue...it is not possible to design a 1/4 wavelength transformer by turning it into distributed RLC's. Any model that tries that does not consider the poynting vector analysis...direction of power flow.

All oscillators bandy back and forth between two energy storage devices...pendulums swap kinetic and potential.tuning forks swap velocity with energy stored via bending....LC resonance swaps inductance and capacitance...transmission line resonators swap energy stored in two different propagation directions..pipe organs work the same way, just acoustically..

It cannot be modelled by RLC analysis..

I haven't worked on the house this week since I am working here. I am hoping to paint next week and close in the room the following week so I can move all my test gear in there
(if the wife lets me)

I'm doin the same, just in the basement..

Cheers, John