Jneutron, please explain your comment from Audioholics. [Archive] - Audio & Video Forums

PDA

View Full Version : Jneutron, please explain your comment from Audioholics.



Tony_Montana
01-28-2004, 05:18 PM
I was browsing Audioholic's cable forum, and I ran cross this post from you:

http://www.audioholics.com/cgi-bin/ikonboard/ikonboard.cgi?s=40184d084485ffff;act=ST;f=4;t=1474

You said in your post:"Coupla decades ago, Bell labs found that the wire inductance in conjunction with the insulation's frequency varying dielectric constant, caused a varying propagation speed through the audio band, making for distortion..they figured it out, and came up with equations to relate it all."

Could you expand on your comment regarding this issue. I didn't know Dielectric constant (absorbtivity) varied with frequency.

For example, Air is a dielectric which have dielectric constant of 1. If the dielectric constant will vary with frequency- then how can you explain white light which is a composite of a number of "different" frequencies all traveling at the same speed thru the air?

Also how would variability of Dielectric Constant effect cable impedance (like Coax) over a wide range of frequency(s)?

jneutron
01-29-2004, 09:20 AM
I was browsing Audioholic's cable forum, and I ran cross this post from you:

http://www.audioholics.com/cgi-bin/ikonboard/ikonboard.cgi?s=40184d084485ffff;act=ST;f=4;t=1474

You said in your post:"Coupla decades ago, Bell labs found that the wire inductance in conjunction with the insulation's frequency varying dielectric constant, caused a varying propagation speed through the audio band, making for distortion..they figured it out, and came up with equations to relate it all."

Could you expand on your comment regarding this issue. I didn't know Dielectric constant (absorbtivity) varied with frequency.

For example, Air is a dielectric which have dielectric constant of 1. If the dielectric constant will vary with frequency- then how can you explain white light which is a composite of a number of "different" frequencies all traveling at the same speed thru the air?

Also how would variability of Dielectric Constant effect cable impedance (like Coax) over a wide range of frequency(s)?

Hi Tony

The dielectric constant of materials varies with frequency...not vacuum's.

I couldn't find the book from which I got that quote. However, from Becker:[Electromagnetic fields and Interactions]:page 260: "both damping and wave velocity will depend upon the frequency, producing a distortion of the signals (of speech, for example) being carried on the conductor."

For zero distortion, (alpha)/omega = sqr(LC)

alpha being the real component of the propagation constant:

gamma=alpha minus j (square root of -1) times Beta.

What it means, is given a real resistance, capacitance, and inductance of the wire pair, there is a sweet spot where no long line distortion occurs.

Unfortunately, that is selected at one frequency (w=5000exp s-1). Dielectrics with frequency varying coefficients will cause the sweet spot to be different for different frequencies, actually eliminating the "sweet spot"

If you peruse JR's site, he mentions the dielectric frequency dependence, and may even link to the belden site where it is stated..


Dielectric constant directly affects capacitance...double one, other follows..so if the dielectric value changes with frequency, the characteristic impedance of the line changes...following Z=sqr(L/C).

Cheers, John

skeptic
01-29-2004, 09:28 AM
Hi Tony

The dielectric constant of materials varies with frequency...not vacuum's.

I couldn't find the book from which I got that quote. However, from Becker:[Electromagnetic fields and Interactions]:page 260: "both damping and wave velocity will depend upon the frequency, producing a distortion of the signals (of speech, for example) being carried on the conductor."

For zero distortion, (alpha)/omega = sqr(LC)

alpha being the real component of the propagation constant:

gamma=alpha minus j (square root of -1) times Beta.

What it means, is given a real resistance, capacitance, and inductance of the wire pair, there is a sweet spot where no long line distortion occurs.

Unfortunately, that is selected at one frequency (w=5000exp s-1). Dielectrics with frequency varying coefficients will cause the sweet spot to be different for different frequencies, actually eliminating the "sweet spot"

If you peruse JR's site, he mentions the dielectric frequency dependence, and may even link to the belden site where it is stated..


Dielectric constant directly affects capacitance...double one, other follows..so if the dielectric value changes with frequency, the characteristic impedance of the line changes...following Z=sqr(L/C).

Cheers, John

How do you explain the phenomenon that when all telephone calls were transmitted over wire and through repeater amplifiers, prior to microwaves, satellites, fiber optics, a coast to coast telephone call would travel through thousands of miles of wire and if there were this distortion to any significant degree, the signal at the receiving end would have been completely unintelligable but it wasn't? What is the typical degree of distortion and how does that relate to the threshhold of hearing? Is this just another example of finding a theoretical problem and blowing it far out of proportion to anything applicable to audio systems?

jneutron
01-29-2004, 10:27 AM
How do you explain the phenomenon that when all telephone calls were transmitted over wire and through repeater amplifiers, prior to microwaves, satellites, fiber optics, a coast to coast telephone call would travel through thousands of miles of wire and if there were this distortion to any significant degree, the signal at the receiving end would have been completely unintelligable but it wasn't?

It is exactly BECAUSE of that application that the research was done to understand the relationship between resistivity, capacitance per unit length, inductance per unit length, and how to combine the three to prevent velocity based distortion.. Before that research, long line transmission was not possible to do well. The example Becker gives has the signal decaying to 1/e at 22 kilometers. Quote:" For cases of minimal damping (R{much less than}wL), the transmission line is distortion free, as the wave velocity is independent of frequency".

So, to answer what I think your question is... the research guys at Bell made it possible..



What is the typical degree of distortion and how does that relate to the threshhold of hearing?

I have no idea..but apparently the distortion was bad enough that the Bell Labs guys had to figure out how the wires were doing it, to enable long line voice fidelity.


Is this just another example of finding a theoretical problem and blowing it far out of proportion to anything applicable to audio systems?

If you choose to make it so, that is your perogative..

I answered a question posed by Tony..if you choose to twist the entire thread around to make it appear that I am "blowing something far out of proportion", then perhaps you could show us exactly where I did so..

In any case...it would appear you have a problem with Becker...not me...

If you wish to e-mail him to tell him he is blowing things out of proportion, you may do so..

Becker, Richard, 1887-1955

Perhaps Dover publications can forward any e-mails to him...I personally prefer to wait at least another 50 years before I am able to talk to him face to face.

Cheers, John

skeptic
01-29-2004, 10:36 AM
"Becker, Richard, 1887-1955"

Damn, isn't it just like those Bell Labs guys to not be around when you need them to answer a straight question!

jneutron
01-29-2004, 10:46 AM
"Becker, Richard, 1887-1955"

Damn, isn't it just like those Bell Labs guys to not be around when you need them to answer a straight question!

I'm rollin on the ground...my sides are hurtin...

For meters of wire, it's a simple thing to calculate the frequency dependence of the wave velocity..so it's really easy to figure how much distortion can happen..

Worst possible case calculation for audio: dielectric ratio of 100 to 1 (which is waaaaay too high a ratio)

Fastest audio signal: say, 95% SOL...(speed of light)

Slowest: what, 5%? 10%??
Time of arrival difference in an audio system? who knows, who cares...5% is 15 million meters per second, one microsecond for 15 meters. I don't think audio cares...

Cheers, John

skeptic
01-29-2004, 11:27 AM
"one microsecond for 15 meters. I don't think audio cares..."

What do you mean audio doesn't care? John Curl cares. He can measure and hear the difference between -120 and -135 db of harmonic distortion of the seventh harmonic of 5 khz. Jon Risch cares. He hears digitial jitter less than that. Especially when he forgets to drink decaf and has a cup of regular java instead.

What kind of audiophile are you if you can't hear a difference of one microsecond? Mtrycrafts is in the back room right now setting up a DBT just to see how good you really are.

jneutron
01-29-2004, 11:46 AM
John Curl cares. He can measure and hear the difference between -120 and -135 db of harmonic distortion of the seventh harmonic of 5 khz.

Yah...did seem that way...

He actually didn't say that he could hear that far down, but that the measurements CORRELATED to what he believes to be the sound of various interconnects.

But...He did approach the issue in a reasonably scientific way..

He went to various lengths to make sure that it was actually the cables that were repeating, and not a weird non repeatable fashion. So, he did SEEM to eliminate the 1700 as the cause...his incorrect assumption was that the machine was perfect.

Now, with at least 3 others repeating the tests on AP machines, he seems to have backed off the cables are non linear stance.

But, that leaves a very open, IMO, important question...The 1700 is DESIGNED to not interact with the cables being tested, to a low level..and yet, at those very low levels being reported, it was. Yes, the equipment was being used below it's guaranteed specifications, but I'm not worried about that...I worry about the repeatability..

There is some, as yet, unexplained interaction between the cable construction, and a machine designed not to interact..

Can that interaction be larger when audio equipment not necessarily designed for lack of interaction(how can you, what equipment will be used?)is being connected?

Of course, that interaction, if audible, should be measureable..I don't think anybody has seen that..


Jon Risch cares. .
He's a different story. He is a guru who HAS to be correct all the time..regardless of what he says..nuff said.

Curl's tests?.....I'll talk of those..JR...doesn't warrant discussion..

Cheers, John

skeptic
01-29-2004, 01:59 PM
I can only hear things that are there. He on the other hand can hear thing that aren't!

Tony_Montana
01-29-2004, 02:11 PM
Thanks Jneutron (and Skeptic) for comments regarding this issue.

You said:"What it means, is given a real resistance, capacitance, and inductance of the wire pair, there is a sweet spot where no long line distortion occurs. Unfortunately, that is selected at one frequency (w=5000exp s-1)."

Would velocity distortion be any way related to cable's resonance properties. For example in antennas, for every frequency, there is specific length of wire which will show resonance properties (at full, half or quater wave length).

So is there any relation between wire resonance (capacitance and inductance are equilibrium) and frequency velocity distortion thru a wire?

mtrycraft
01-29-2004, 08:48 PM
"Becker, Richard, 1887-1955"

Damn, isn't it just like those Bell Labs guys to not be around when you need them to answer a straight question!


No, skep, that is why we have John Edwards and Sylvia Brown to contact them for answers :)

mtrycraft
01-29-2004, 09:09 PM
I'm rollin on the ground...my sides are hurtin...

For meters of wire, it's a simple thing to calculate the frequency dependence of the wave velocity..so it's really easy to figure how much distortion can happen..

Worst possible case calculation for audio: dielectric ratio of 100 to 1 (which is waaaaay too high a ratio)

Fastest audio signal: say, 95% SOL...(speed of light)

Slowest: what, 5%? 10%??
Time of arrival difference in an audio system? who knows, who cares...5% is 15 million meters per second, one microsecond for 15 meters. I don't think audio cares...

Cheers, John

Stephen H. Lampen of Belden has a book out, "Wire, Cable, and Fiber Optics for Video & Audio Engineers," McGraw Hill, 1997.

Here are some DC:

rubber 3-6
EPDM 2.5-6
PVC 3.5-5.5 Vp 45%-57%
teflon 2.1 Vp 69%
polye 2.3
polyp 2.23 Vp 67%

Time delay in cables, as some manufacturers prefer, is measured in nanoseconds/ft.

Vp(Velocity of propagation) = 100/sqrt of DC

Now it is easy to calculate propagation in these cables, close to 1ft/nsec.

mtrycraft
01-29-2004, 09:13 PM
[QUOTE=skepticMtrycrafts is in the back room right now setting up a DBT just to see how good you really are.[/QUOTE]


I am ready :) And pctower is here to make sure we follow Leventhal's tables, so, it will be a long night, very long :)

mtrycraft
01-29-2004, 09:19 PM
But, that leaves a very open, IMO, important question...The 1700 is DESIGNED to not interact with the cables being tested, to a low level..and yet, at those very low levels being reported, it was. Yes, the equipment was being used below it's guaranteed specifications, but I'm not worried about that...I worry about the repeatability..

There is some, as yet, unexplained interaction between the cable construction, and a machine designed not to interact..

Can that interaction be larger when audio equipment not necessarily designed for lack of interaction(how can you, what equipment will be used?)is being connected?
Cheers, John


But John modified his 1700. So his modification is also suspect, or a good possible contributor to the problem.

mtrycraft
01-29-2004, 09:21 PM
"one microsecond for 15 meters. I don't think audio cares...".


That is closer to 50 +/- nanoseconds

skeptic
01-30-2004, 04:39 AM
I've been trying to contact Jackie for two years now. Her voice mail still works but she never retruns my calls. I know she's there because her friends will see to it that she will be the last one to close the door. At least when she was at Bellcore and she wasn't in her office, I knew I could always find her at the mall with the best sales at that moment.

Most of the regulars have VRIP'd or been FMP'd. I'll bet Murry Hill is a ghost town.

pctower
01-30-2004, 05:08 AM
I am ready :) And pctower is here to make sure we follow Leventhal's tables, so, it will be a long night, very long :)

Well it's about time we finally got around to some real science around here.

mtrycraft
01-30-2004, 08:34 PM
Well it's about time we finally got around to some real science around here.


Yes, yes, good science takes time :)

FLZapped
01-31-2004, 07:02 AM
But, that leaves a very open, IMO, important question...The 1700 is DESIGNED to not interact with the cables being tested, to a low level..and yet, at those very low levels being reported, it was. Yes, the equipment was being used below it's guaranteed specifications, but I'm not worried about that...I worry about the repeatability..

There is some, as yet, unexplained interaction between the cable construction, and a machine designed not to interact..

Cheers, John

Hi John,

It may not so much be the cable, but the load. If I remember correctly, the artifacts went up with decreasing load resistance.

-Bruce

Chuck
01-31-2004, 05:13 PM
Bell Labs guys had to figure out how the wires were doing it, to enable long line voice fidelity.

Hi John,

You said something important in your first post (on Audioholics.com) that is worth keeping in mind, and is directly related to what you say in the quoted comment above. The operative words here are "long line," and 10-feet of wire isn't what Bell Labs was concerned with. As the wire-run gets shorter, the problems are diminished, and at some point they become non-issues, at least as far as audibility is concerned. Your earlier post on Audioholics.com clearly stated that the effects with 10-feet of wire were insignificant:



Ten feet? for audio freq's, it's not a problem.

Nuf said. :)

jneutron
02-02-2004, 07:48 AM
That is closer to 50 +/- nanoseconds

15 meters is 45 feet.
at SOL, that is about 45 nanoseconds..as you state..

5% of speed of light is 20 times the delay, or 900 nano's, close to a microsecond. I was talking about the difference in time a 5% speed would give..sorry I wasn't clearer..

Cheers, John

jneutron
02-02-2004, 07:54 AM
Thanks Jneutron (and Skeptic) for comments regarding this issue.

You said:"What it means, is given a real resistance, capacitance, and inductance of the wire pair, there is a sweet spot where no long line distortion occurs. Unfortunately, that is selected at one frequency (w=5000exp s-1)."

Would velocity distortion be any way related to cable's resonance properties. For example in antennas, for every frequency, there is specific length of wire which will show resonance properties (at full, half or quater wave length).

So is there any relation between wire resonance (capacitance and inductance are equilibrium) and frequency velocity distortion thru a wire?

No. the velocity distortion is related to how each frequency propagates down the line.

Specific lengths will resonate only by virtue of reflections at each end, like water in a bathtub.

You are mixing terms..cap and induct at equilibrium is an interesting one.. In a transmission line, C and L form the basis for characteristic impedance and prop velocity..not resonance...an infinite transmission line does not resonate..

Cheers, John

jneutron
02-02-2004, 07:57 AM
But John modified his 1700. So his modification is also suspect, or a good possible contributor to the problem.

Absolutely. As is the geometric layout of the machine. Whatever the cause of the low level harmonics, he reported a CABLE specific, repeatable change..

I'm not concerned about the levels, but the repeatablility..

Cheers, john

jneutron
02-02-2004, 07:58 AM
Hi John,

It may not so much be the cable, but the load. If I remember correctly, the artifacts went up with decreasing load resistance.

-Bruce
I think everything is playing in.. I'm concerned with the repeatability vs cable.

Cheers, John

FLZapped
02-02-2004, 09:26 AM
I think everything is playing in.. I'm concerned with the repeatability vs cable.

Cheers, John


Agreed, but it's always nice to know how your stuff behaves......or misbehaves.

-Bruce :D

Worf101
02-02-2004, 12:58 PM
"one microsecond for 15 meters. I don't think audio cares..."

What do you mean audio doesn't care? John Curl cares. He can measure and hear the difference between -120 and -135 db of harmonic distortion of the seventh harmonic of 5 khz. Jon Risch cares. He hears digitial jitter less than that. Especially when he forgets to drink decaf and has a cup of regular java instead.

What kind of audiophile are you if you can't hear a difference of one microsecond? Mtrycrafts is in the back room right now setting up a DBT just to see how good you really are.

Man you are just too too cruel. This had me laughin my head off... thanks... :D

mtrycraft
02-02-2004, 09:52 PM
I'm not concerned about the levels, but the repeatablility..
Cheers, john


And, since he uses the same setup but changes cables and repeats the same cables, why would it not repeat. He didn't eliminate any of th epossible source of his trouble as he is too bull headed to think he made a mistake.

jneutron
02-03-2004, 07:38 AM
And, since he uses the same setup but changes cables and repeats the same cables, why would it not repeat. He didn't eliminate any of th epossible source of his trouble as he is too bull headed to think he made a mistake.

If it were solely the 1700, one would expect the same spectra within each cable. Since he found that it tracked the cables, he assumed it was just the cables...I do not..I assume that an interaction between cable and 1700 caused the repeatability.

I agree he did not isolate the cause..

Bull headed? perhaps...but, he was bull headed enough to stand his ground when others said it was bogus...he was bull headed enough to use a methodology to determine that the effect was indeed VERY correlatable to cable, and (presumeably) not the 1700.

His only error (methodological) was that he assumed there was no cable-1700 interaction that could be responsible for the spectra differences..

One error...big deal...I can do that in five minutes...

Cheers, John