New title - Why do you need to try high end cables? - Page 9

That *sounds* plausible - is there a way to test that jnjn? 

Also - would it be audible, and is it an issue when comparing normally constructed ICs with one another (e.g. if the only difference between two IC's is the primary conductor's metal)? 

Quote:

Originally Posted by liamstrain 

That *sounds* plausible - is there a way to test that jnjn? 

 

Also - would it be audible, and is it an issue when comparing normally constructed ICs with one another (e.g. if the only difference between two IC's is the primary conductor's metal)? 

It is by no means an easy test.  It would require setting up 4 instrumentation amps, (balanced differential inputs), with two from the source, two at the final load.  Subtract each input from it's corresponding output to null.

Once nulled, try physical manipulation of the cables.  If there are two line cords involved, they too will be suspect.

Beyond simple noise and supply AC current pulses modulated by the music power amplitude, one needs to evaluate the lateralization parametrics responsible for image differential placement.   I would look down to at least a tenth of a dB interchannel intensity difference, and 5 uSec interchannel timing difference.

Getting a true null is not going to be easy.  Remember, the biggest issue will be interpretation of any timing shifts, a simple differential produces the difference, but does not directly indicate if it is amplitude or time caused.  It would be necessary to also have a sync signal to determine which it is.

The loop paths are dependent on the impedances of the path.  At DC for example, if there are two safety grounds involved, the signal return current will travel along the line cord ground exclusively.  At roughly a kilohertz, the switchover to shielded operation will be happening.

Cheers, jn

ps. and yah steve, "friends don't let friends do AC.."

pps.  how do we put drawings in here?

Ah, you remember. Did that one many moons ago.

If you look at the editing icons, you'll see an "image" next to a strip of film that's next to a paperclip. Click on the "image" and you'll be prompted to either supply a URL for the image you want to include, or you can select it off your local drive.

se

Quote:

Originally Posted by jnjn 

It is by no means an easy test.  It would require setting up 4 instrumentation amps, (balanced differential inputs), with two from the source, two at the final load.  Subtract each input from it's corresponding output to null.

 

Once nulled, try physical manipulation of the cables.  If there are two line cords involved, they too will be suspect.

Beyond simple noise and supply AC current pulses modulated by the music power amplitude, one needs to evaluate the lateralization parametrics responsible for image differential placement.   I would look down to at least a tenth of a dB interchannel intensity difference, and 5 uSec interchannel timing difference.

 

Getting a true null is not going to be easy.  Remember, the biggest issue will be interpretation of any timing shifts, a simple differential produces the difference, but does not directly indicate if it is amplitude or time caused.  It would be necessary to also have a sync signal to determine which it is.

 

The loop paths are dependent on the impedances of the path.  At DC for example, if there are two safety grounds involved, the signal return current will travel along the line cord ground exclusively.  At roughly a kilohertz, the switchover to shielded operation will be happening.

 

And with that, I'm out of my depth. But thank you.  :D

It sounds like we are dealing with relatively minor influences here though (tenths of a db). Do you think this would be audible? Also, can you address the last part of my post, if you know - regarding similarly constructed IC's with only changes to the conductor metal? 

Quote:

Originally Posted by liamstrain 

And with that, I'm out of my depth. But thank you.  :D

 

It sounds like we are dealing with relatively minor influences here though (tenths of a db). Do you think this would be audible? Also, can you address the last part of my post, if you know - regarding similarly constructed IC's with only changes to the conductor metal? 

 

Don't forget, while .1 dB by itself is not distinguishable, one must consider an interchannel difference, not an absolute one.  What we are trying to discern is image placement relative to other images.  Most researchers confuse the two.  I've not stated that .1dB is the effect level, just that I would recommend the measurement tools be capable of that level of accuracy and precision in addition to the 5 uSec entity (which I would recommend 100 nSec precision).  I am not interested in equipment which has a resolution equal to the measured entity, but rather, at least an order of magnitude better if possible.

Ah, the last part..sorry.

The metals used can alter the permeability, resistivity, and geometry. If a magnetic wire is used, the internal inductance becomes much higher which affects the cable's characteristic impedance.  This alters the breakpoint frequency where current will share 50/50 with the line cord neutral/ground, as well as with the other IC.  Granted, it's only 15 nH per foot, but a high capacitance cable will be affected more.

The resistivity of the shield determines the amount of current which will flow in the ground loop, and will again affect the 50/50 breakpoint frequency.

The geometry of the shield determines the inductance of a coax...a spiral shield of course, higher L and higher R.

Honestly, I wouldn't worry about a choice of silver over copper, it's trivial.

Worry about the equipment...that is where the real problem arises.  Unbalanced input/output audio equipment does not meet EMC design at all..and balanced is a tad better, but not much.

A well designed system will be invariant to the choice of IC's and PC's.  Any alteration whatsoever in a system as a result of wire changes indicates that the equipment has problems.

Personally, I use the IC's and PC's that come free with the equipment.  If I find any issues with my wire selection, I generally fix the equipment or re-layout the wires.

ps..added a jpeg showing the calculation of the ground path reactance for a simple one IC system.  The electrical diag shows the basic elements the return current has to choose from to get back to the source.

Cheers, jn
 

Awesome. That's about where I had come out myself - just without the electrical engineering depth of knowledge to back up the why's. 

Thanks again for your detailed and helpful answers. 

My head just exploded.  Thanks for the post jnjn.  Mostly over my head but I will pretend I understood some of it.  

Any chance of a critique of Shunyata white paper testing methodology?

You mean other than their "measurements" not having any particular relevance to the real world?

It's just a dog and pony show. If they wanted to demonstrate in any meaningful fashion that their power cords are more beneficial than a stock power cord, they'd show their effect on the output of a typical linear power supply. But they don't. Can't imagine why.

se

Quote:

Originally Posted by Steve Eddy 

 

You mean other than their "measurements" not having any particular relevance to the real world?

 

It's just a dog and pony show. If they wanted to demonstrate in any meaningful fashion that their power cords are more beneficial than a stock power cord, they'd show their effect on the output of a typical linear power supply. But they don't. Can't imagine why.

 

se

 

I had a feeling their results might be dodgy in terms of their assumptions and test setup eg. using low voltage and high currrent to make differences in inductance and resistance larger, but I'm no electronics guru.

Quote:

Originally Posted by drez 

My head just exploded.  Thanks for the post jnjn.  Mostly over my head but I will pretend I understood some of it.  

 

Any chance of a critique of Shunyata white paper testing methodology?

 

 

 

Shunyata does not consider the distinction between causation and correlation.  They are working and selling under the incorrect belief that the high speed transient response of the line cord directly affects the system being modified.

As such, their explanations are not consistent with reality.

That said, the inductance and geometry of a line cord can affect a system.  But not how they say it does.

1.  Unbalanced connected equipment fails the first tenet of Electromagnetic Compatibility theory.  By design, there is inadequate control of the currents within the system.

2.  Equipment which is poorly designed with respect to EMC will be susceptible to ground loop problems.  Not just line frequency, but the entire audio band.

3.  Any modifications which change the agressor, the victim, or the coupling mechanism of the system will by nature, alter the characteristics of the system.

To have EMC problems requires an agressor, a victim, and a coupling mechanism.  A line cord can be an agressor via the external magnetic field of the line cord.  Lower inductance cords will have smaller external magnetic fields. 

The line cord is also part of the coupling mechanism.  Remember, the IC loop doesn't control the path of all the audio current provided by the source component.  And the power cord is the higher conductivity leg of that complex IC loop system.

I tried to add a drawing, but the system puked on a pdf..I'll have to find the source and make a jpg.

Cheers, jn

Here is a pic showing the ground loop.  Note that a source is also capable of picking up problems from the loop, not just the cable nor amp.  That is represented by the pink area in the source chassis.  The center cable error source is generically the IR drop of the loop current through the IC shield.  The green shaded is the amplifier loop error.  The blue is the generic external field loop trapping area where the ground currents typically are created.  Note however, that this does not detail how an amp can create ground loop currents, nor specifically how the comps convert that current into error signals which enter the audio stream.  Bill Whitlock covers only the green "pin 1" problem with respect to IR drop only, but that is not the entire story.

Here is another pic, showing how to test a specific line cord as an agressor.  Using the swept sine source, look for induced current in the loop formed by the two line cords.

cheers, jn