OK Scott, there is more measured distortion in audio cables when used one direction vs the other. If you are ever in town I'll be glad to demo it for you. Check with JJ if you want a first hand account of the demo. ES
Love to, I would be glad to sit down with an experiment and then deduce what is really being measured. That is my specialty, I only hope you will accept an answer when it is presented. If your experiment shows superposition is violated it is wrong plain and simple.
I ask ALL the questions and don't give up.
No.
If cables are directional so would PCB traces and especially flexible circuits (plenty of grain boundries, depending whether the copper is electrodeposited, wrought or electroplated).
The only time I have heard anything to do with cable directivity is on Audio sites, no other area of electronics even considers the posibility that a wire will only pass the signal in one direction. As electons dont travel very far or very fast (8.4cm per hour), I cant see HOW a cable can be directional, when it is effectively acting as a wave guide. What are the mechanisms, how can ac current be used etc etc.
Of all the audiophile myths, this is the most stupid one IMO. Arghhh!!!!
If cables are directional so would PCB traces and especially flexible circuits (plenty of grain boundries, depending whether the copper is electrodeposited, wrought or electroplated).
The only time I have heard anything to do with cable directivity is on Audio sites, no other area of electronics even considers the posibility that a wire will only pass the signal in one direction. As electons dont travel very far or very fast (8.4cm per hour), I cant see HOW a cable can be directional, when it is effectively acting as a wave guide. What are the mechanisms, how can ac current be used etc etc.
Of all the audiophile myths, this is the most stupid one IMO. Arghhh!!!!
If distortion is present then superposition is automatically violated, as superposition only works for linear networks. Reciprocity, maybe?
It is possible to make a passive lowish loss two-port circuit which is directional e.g. an SWR meter. It ought to be possible to take this idea and add some mechanism to reflect distortion back onto the line. Therefore a directional low-loss distortion producer does not appear to break any basic circuit theorems. (I am thinking aloud here, or woffling - depends on your point of view.) It is quite a big step from that to say that it can be done simply by putting a long thin lump of copper next to a long thin lump of PVC (or some other dielectric).
It is possible to make a passive lowish loss two-port circuit which is directional e.g. an SWR meter. It ought to be possible to take this idea and add some mechanism to reflect distortion back onto the line. Therefore a directional low-loss distortion producer does not appear to break any basic circuit theorems. (I am thinking aloud here, or woffling - depends on your point of view.) It is quite a big step from that to say that it can be done simply by putting a long thin lump of copper next to a long thin lump of PVC (or some other dielectric).
Funny the RF guys regularly will check cables and then use them in the direction that shows lowest loss.
Once you go past basic circuit theory all sorts of possibilities show up.
But try Audioxpress 11/09 for more details.
In RF there may be a real effect. No cable, and no termination, is perfect. If a cable happens to wander from around 49ohms impedance at one end to around 51ohms at the other, and the terminations happen to match then you will get no reflections one way round and small reflections the other way round. In this sense the cable is directional. In addition, connections to plugs and sockets are not perfect so a minor mismatch in one can partially cancel a minor mismatch in the other.
The same situation does not arise in audio, as audio cables are severely mismatched at both ends.
The same situation does not arise in audio, as audio cables are severely mismatched at both ends.
Now, that IS interesting!
That's intriguing!
Yo, detractors: aside from what we want to think or believe is audible or true, when something has been measured and is repeatable, it's time to ask more questions, not condemn the information we've been presented with.
Simon: What kind of numbers is your testing showing?
And what kind of relative difference between the numbers?
For example if we're seeing 0.00010 vs 0.00007, one is about 30% more distortion than the other.
What do you think could be the causes of this difference?
Here's are some of the things I'm thinking could be part of this issue:
- Low level contact rectification, more CR on one connector than the other?
- Electrostatic charge on the dielectric making the connector look like it is a very small polarized capacitor? Any accurate way of testing for micro-amounts of DC?
- Mismatched impedances...one of the previous posters said this couldn't possibly apply to audio as BOTH terminations are mismatched. Uh... if a slight mismatch is compounded with other larger mismatches, there is a higher probability of distortion, not a lower one. The probability is that the mis-terminations would not be symmetrical, yes?
Please: Don't come at me about the audibility or non-audibility of such, I make no claims pro or con. I'm looking for technical answers, not arguments.
I'm in the middle of making some interconnects, and jus wanna make cables that measure better and sound really good, that's all.
If you don't believe it's audible, that's not useful information for me, so please just keep it to yourself. No dogmas, please.
Some logical inferences:
If it's routine practice for RF engineers to change the cable orientation to optimize reflections, then it's clear that there is asymmetry somewhere in the cable, and that it affects performance.
If Simon's measurements are showing differences in distortion, then it's clear there is asymmetry somewhere in the cable, and that it may indeed affect performance.
Now this:
When they are correctly done, and repeatable, measurements present us with valid data. Then it is up to us to correctly interpret that data.
To blindly reject measured data because we choose to believe "that just can't be"... well, so much for objectivity or science.
Actual differences in distortion?
That's intriguing!
Yo, detractors: aside from what we want to think or believe is audible or true, when something has been measured and is repeatable, it's time to ask more questions, not condemn the information we've been presented with.
Simon: What kind of numbers is your testing showing?
And what kind of relative difference between the numbers?
For example if we're seeing 0.00010 vs 0.00007, one is about 30% more distortion than the other.
What do you think could be the causes of this difference?
Here's are some of the things I'm thinking could be part of this issue:
- Low level contact rectification, more CR on one connector than the other?
- Electrostatic charge on the dielectric making the connector look like it is a very small polarized capacitor? Any accurate way of testing for micro-amounts of DC?
- Mismatched impedances...one of the previous posters said this couldn't possibly apply to audio as BOTH terminations are mismatched. Uh... if a slight mismatch is compounded with other larger mismatches, there is a higher probability of distortion, not a lower one. The probability is that the mis-terminations would not be symmetrical, yes?
Please: Don't come at me about the audibility or non-audibility of such, I make no claims pro or con. I'm looking for technical answers, not arguments.
I'm in the middle of making some interconnects, and jus wanna make cables that measure better and sound really good, that's all.
If you don't believe it's audible, that's not useful information for me, so please just keep it to yourself. No dogmas, please.
Some logical inferences:
If it's routine practice for RF engineers to change the cable orientation to optimize reflections, then it's clear that there is asymmetry somewhere in the cable, and that it affects performance.
If Simon's measurements are showing differences in distortion, then it's clear there is asymmetry somewhere in the cable, and that it may indeed affect performance.
Now this:
When they are correctly done, and repeatable, measurements present us with valid data. Then it is up to us to correctly interpret that data.
To blindly reject measured data because we choose to believe "that just can't be"... well, so much for objectivity or science.
Last edited:
Is this available online?
Thanks Simon. Is this or a summary of it available anywhere as online free info?
Only some of the graphs are at the AX website. If you find it online it is copywrighted, let me know.
I seem to have misplaced it, I am perfectly willing to repeat your tests. I just hope you don't mind if I bite into this and get some perfectly good answers.
EDIT Found it, those were the pictures with so much unexplained noise, line, artifacts, etc. that I would not make any conclusions about anything. I would gladly repeat this as an experiment (in a screen room if need be) and isolate the "distortion" if any.
Last edited:
It is on my to do list to make a bunch of the testers as the heart of it is not too complicated.
I think we are in agreement that superposition does not hold for nonlinear systems?
Do it!
Thanks in advance!
That would be valuable! especially if you follow simon7000's test protocol, and could then publish your test procedures and the numbers on this thread.
Thanks in advance!
Actual differences in distortion?
- Low level contact rectification, more CR on one connector than the other?
- Electrostatic charge on the dielectric making the connector look like it is a very small polarized capacitor? Any accurate way of testing for micro-amounts of DC?
- Mismatched impedances...one of the previous posters said this couldn't possibly apply to audio as BOTH terminations are mismatched. Uh... if a slight mismatch is compounded with other larger mismatches, there is a higher probability of distortion, not a lower one. The probability is that the mis-terminations would not be symmetrical, yes?
As of yet these postulates keep reappearing with no experiments done to isolate any of them.
As for actually hearing these -130dB distortions, folks here have tried to get even one person from each side in the same room. Nope, all failed. It's just like the Hatfields and McCoys.
If it can be measured reliably and repeatably (which means at different places too, ideally with somewhat different setups so it is the cable being measured and not the test equipment) then I agree it is time to ask more questions. My guess is that plugs and sockets will be the issue, as things happen at interfaces which don't happen so easily in bulk materials.Jack Caldwell said:
However, and I am not just being awkward, we must remember that sometimes in science it turned out that the experiments were wrong not the theory.
In order to distinguish direction of power flow (which essentially is what is being claimed) you need to combine voltage and current, because neither on its own indicates direction. The insulation would be the prime candidate for this, but I think most common dielectrics have no magnetic effect. There are exotic materials (usually ceramics) which react to both, and you can do clever things with biased ferromagnets (e.g. microwave circulators), but cable materials are generally simpler than this.
I don't think you said that but left it unsaid.
I think we both know that copper has a temperature coefficient (3930 ppm/K) to it's resistance. So at some level even a copper wire is a nonlinear resistor. (Now if one wishes to argue that is what is being heard in cable comparisons we may both agree that is probably not the cause... special really stupid examples excepted.)
We both would expect the wire's thermal distortion due to tempco to be the same with respect to direction. That would leave connector non-concentricity or similar issues to change the mating resistance on each end. I think we can agree that could be measured as a directional interconnect. (Again if that becomes enough of an issue to be picked up by ear is a different issue.)
So the experiment I did not conduct was to unsolder the cables under test and swap the plugs. So after I finish my current to do list I may revisit that.
There would have to be a significant change in mating resistance. To first order, the mating resistances are merely in series with each other so direction does not matter. To second order it does matter because the cable capacitance sits between them. However, mating resistance will normally be many orders of magnitude below the source resistance so swamped.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.