cables for esl speakers

glorocks

Member
2007-07-05 5:53 pm
I'm hoping some of you may want to discuss your experiences in finding or building optimized amp to ESL cables. I've read quite a number of white papers on the subject, most notably those of Allen Wright and Rodger Sanders.
Sanders' paper made a lot of sense on a number of levels, and Wright's discussion of skin effect was to me, compelling.
ESL's pose a very unique load. as we know. Stray inductance seems to be the biggest problem, followed by excessive capacitance.
Sanders maintains that cables exhibiting a higher resistive impedance are desireable. To me this makes sense, as ESL's are essentially voltage rather than current driven. To me, this suggests thin conductors- very thin.
Inductance occurs whenever signal carrying conductors are in proximity of each other as I understand it. It can be cancelled to a degree if the opposing conductors cross each other repetitevly as in a braid or Litz confiuration according to several engineers I have spoken to.
Another area I would like to explore is the treatment of the interconnect as a transmission line. Bell Labs established years ago that a properly designed line can minimumize a number of distortion components.
So, friends can we talk?
 
Last edited:
Hi,

as usual with cables there´s no general recommendation of what´s best.
As for example the cited increased resistivity. The upper bandwidth limiting resonance resulting from the stray capacitance of the audio tranny and the panel capacitance requires some damping. The value of the damping resistor can vary between a fraction of ohms to a few ohms. If You use a low output impedance amplifier, like most fed back transistors You certainly need a dedicated resistor to perform the task, because there´s no cable with such a high resistivity. If You use a no-feedback tube amp You might not even need a resistor at all, because of the higher resistivity of the tube amp´s secondary winding resistance. In any way can You tune bandwidth and frequency response of the panel with this resistance. But You can´t recommend a certain value as a onefitsall-value. As with most speaker cables it should be a low resistivity cable to start with.
Since the amp typically sees a heavy capacitive load (in the middle to high frequency range) of couple of µF a additional capacitive load of a few hundreds of pF added by the cabling doesn´t change matters much.
I´d rather opt for a low inductance cabling instead.
But then again, there´s no general rule.

jauu
Calvin
 
As Calvin says, for an ESL you can use a cable which trades off parasitic capacitance in favor of low inductance, since a little extra capacitance won't make a difference.

The easiest way to make such a cable is from computer flat-cable. Simply alternate between + and - signals: all even conductors carry +, all odd conductors carry -.

As an added bonus it's possible to run this cable under the carpet if necessary :)

I can post a picture later if desired.

Kenneth
 
Inductance occurs whenever signal carrying conductors are in proximity of each other as I understand it.
No - self inductance increases as the two conductors are moved away from each other. The braiding or Litz configuration is to reduce skin effect on very large diameter wires. Skin effect is not a problem on small diameter ELS cables. Characteristic Impedance is not a factor in speaker length cables.
 
About 25 years ago, I bought 100 feet of 3-wire, 16 gauge, cord made for heavy duty extension power cords, at a yard sale. Each of my tweeters, about 1KV bias, has around 25 feet of it... just in case I move to a big room next.

Can't say as I think (or mic's think) there is any deficiency in my tweeter system. Anybody think I've gone wrong?

On the other hand, when I was working DIY with Dayton-Wright panels, I found it convenient to fish test-lead HV wire for the bias and ordinary wires for the stators through clear plastic tubing. 3-4 KV bias with those panels.

ESL panels are pretty low capacitance so as a high-reactance load, hard to upset them with upstream wires with minuscule losses. Likewise, the whole system south of the power amp is wacky in several frequency-interfering respects which can't easily be assessed.

In short, very low on the list of things worth fussing over, I'd say.
 
Last edited:

glorocks

Member
2007-07-05 5:53 pm
I understand the function of the series resistance. Sheldon Stokes mentioned it in his DIY ESL project, and the data that was provided with my transformers (Plitron) suggested a starting point of about 1 ohm.
I suppose I will have to do an RTA to look for that pesky peak and probably have to use the old substitution method to determine the desired value. Thank you for you contribution-
 
Hi,

I did a little test a while back.

1.5mm^2 standard "lamp" cable:

ESLimp2.jpg


RG213 coax cable off good quality

ESLimp1.jpg


The big differance is in the phaseshift - 20deg differance at 20kHz!

What it does in reality I do'nt know - Can't hear any differance!

[Connected speakers are my DIY ESL-hybrids]

/R
 
Electrostatics generally look like capacitor to amplifiers, need relatively higher current at higher frequencies. But some are more inefficient than others, need more current. My Acoustat 1+1's are in the super-inefficient category (Black Hole). The bigger Acoustats are actually more efficient, the smaller ones less efficient, that's the way it worked. The more inefficient your speakers are, the more series impedances like resistance and self-inductance matter as they will soak up power in a way that alters frequency response. Given that you are already using 16ga or larger wire, self-inductance is probably a more important factor for electrostatics in modifying their frequency response.

I went from 12ft 16g cord to 4ft. Nice difference. Problem is, I'm going to have to go back to 12ft eventually when I get my old amplifier back. So how do I get the same self-inductance in a 12ft wire as I would get in 4ft 16g zipcord?

I think I will do just that. Many different multiconductor designs will work. I'm thinking of getting some variety of Cardas; they use Litz combinations and specify the self-inductance of their different cables. That's nice. But there are many other cheaper options too, and you can test cheap braided/flat wires to see what their self-inductance is.
 
I went from 12ft 16g cord to 4ft. Nice difference. Problem is, I'm going to have to go back to 12ft eventually when I get my old amplifier back. So how do I get the same self-inductance in a 12ft wire as I would get in 4ft 16g zipcord?

Or you could run coaxial cable, those have low ESL. I've had good luck with parallel runs of RG-8. RG-58 is lighter, but a lot more tractable than RG-8.
 
Some other low-inductance designs here:

DIY speaker cable for ESL

2m stacked ribbon with << 0.1uH? Wow that's low. Too bad capacitance rises to 10nF. I noticed you had your << signs turned the wrong way >>.

Would be good to see picture or diagram of your winning design, twisted pair #3 with PTFE dialectric. Where did you get the PTFE insulated silver plated wire? Looks like you got a winner there!

Also glad to see we understand the problem. I can't seem to convince some of my friends that anything matters other than wire gauge, or others that something more is needed than a box of crystals.
 
Hi Charles,

that site is not by me :) You're right though, the << are the wrong way around.

One place where you can get PTFE silver plated wire is Apex Jr.Home Page

Yes, cable inductance is much more important for ESLs than for conventional speakers... 10nF cable capacitance is not much of a problem, it is swamped by the (transformed) panel capacitance.

BTW, see photo below for what I did. The alternating wiring scheme minimizes inductance (I don't have a tool to measure it).

Kenneth
 

Attachments

  • ribboncable.jpg
    ribboncable.jpg
    116.3 KB · Views: 290
Note on methods

snip
I went from 12ft 16g cord to 4ft. Nice difference. Problem is, I'm going to have to go back to 12ft eventually when I get my old amplifier back. So how do I get the same self-inductance in a 12ft wire as I would get in 4ft 16g zipcord?

snip

OK, you can tell from my previous post that I am a skeptic about the influence of cables. Maybe I just am blindly carrying over my reactions from reading nutbar advertizing about expensive cone speaker cables. And I don't mean to appear to be criticizing this particular post, just quoting it as a starting point.

But without an ABX box, here's the kind of method I would use before forming an opinion, let alone posting it. I'd have one speaker of a stereo pair on a 4 foot cord and the other on a 25 foot cord. Playing a mono signal.

Stereo-sense is a pretty sensitive "meter."

Piece of cake to do. Bob's your uncle.
 
Last edited:
There's a very nice analysis of the effects of reactances and resistances in the source impedance driving an ESL in Menno Vanderveen's new book on tube amps. Series resistance can be a major issue in the bass response. With my ESLs (rebuilt Acoustat 1+1), I went the route that Charles did, a 3x12 gauge extension cord cut down. Worked like a charm.
 
There's a very nice analysis of the effects of reactances and resistances in the source impedance driving an ESL in Menno Vanderveen's new book on tube amps. Series resistance can be a major issue in the bass response. With my ESLs (rebuilt Acoustat 1+1), I went the route that Charles did, a 3x12 gauge extension cord cut down. Worked like a charm.

With little prospect of me ever seeing that book, could you do the math of your system for us and show why it could make a difference, please. That is, different from the non-exotic sort of middle-of-the-road solution that you and Charles are adapting.

Are which side of the step-up transformer are we worried about?

Again, maybe it is my hangover from the discussions of cone speaker cables, but over there, there seems to be a bright-line dividing the techno-rationalists math-enabled posters from the golden-eared-subjectivists.

And if it made a difference of phase response at 20kHz or something for ESLs.... where on the list of factors that matter and could readily be addressed, does cable shortcomings fall?
 
Last edited:
How well do these ribbon cables work ? I'm running some Final 'stats which are cabled with 1.2mm transformer wire at the moment. I quite like this wire (as good as any budget / mid price cable I've ever tried) but it's inflexible. So wondering about other options although would ideally like something in a nicer colour than grey !
Looks like these are the 25 strand versions - does it matter which is used ? I've seen cable with 10, 20 and 25 starn (all 28awg).
 
Hi,

we´ve made best experience with these cables NEYTON High-End Audio Cables - Frankfurt LS and the Berlin-LS (in a cross connected low-Ind. version)
Depending on the wiring, You can have a low inductance or low capacitance version, which ever suits Your speakers best. For ESLs a low inductance version is preferrable. Build and material quality is truely awesome (better than the similar but costier Nordost cables). While these are for sure outside a DIYers wallet, it can give hints into the construction details of a well suited ESL cable.

jauu
Calvin
 

alexberg

Member
2009-10-01 2:25 pm
It's much easier than one may think...
two stripes of foil 1 sm wide separated by 1 mm do have AC inductance around few hundred nH per meter (sure if they are forward and return, or carry opposite and equal currents). Wires in the cable Calvin has mentioned could be either connected interleaved or separated... yeilding either low AC inductance or capacitance.
P.S.
"The Self and Mutual Inductances of Linear Conductors"
By Edward B. Rosa, Washington, Sept. 15, 1907