How essential is shielding?

Hi Bryon,

My opinion is that the case for shielding is strongest when it comes to unbalanced interconnects, both line level and phono level.

In the case of unbalanced line level interconnects, signal return currents flow in common with extraneous inter-chassis noise currents which will be present to some degree due to ground-loop effects, as explained in this paper. Since the receiving component cannot distinguish between signal and noise that may be present at its input, the voltages corresponding to those extraneous currents in the return conductor must be minimized. That means minimizing the impedance of that connection path, which in turn will be accomplished by a quality shield, but not necessarily by unshielded configurations. And certainly not by a cable configuration in which the return conductor is a simple piece of wire, which will have substantial impedance at high frequencies due to its inductance, and perhaps also higher impedance at low frequencies due to higher resistance.

As explained in the reference, an important factor in all of that is stray capacitance in the power transformer of each component. Which means that the magnitude of this effect will be system-dependent and unpredictable.

Shielding will also, as others have said, reduce noise pickup caused by rfi/emi effects, which is particularly important in the case of unbalanced interfaces due to their poor common mode noise rejection capability, and in the case of phono level interconnects due to the low signal levels that are involved.

That said, it's always been a mystery to me how, in particular, Anti-Cables unbalanced interconnects, which use a coiled piece of wire as the return conductor, apparently perform very well for many users.

All of the foregoing considerations are pretty much inapplicable to balanced interfaces, as you will realize, and so I would expect shielding to be less important for balanced interconnects.

And I would expect it to be unimportant for speaker cables, with the inter-chassis noise current consideration being inapplicable to a passive speaker, and because sensitivity to emi/rfi effects is lessened by the low impedances and relatively high signal levels that are involved. Although I do recall one thread from some time ago in which an rfi problem was resolved by changing the speaker cables.

Best regards, and happy new year!

-- Al

A comment on the Cardas video linked to above. I don't disagree with anything that was said, but I would point out that the main reason for the dramatic changes in the waveform displayed on the oscilloscope as the cable was moved about is that the other end of the cable was not connected to anything. If it had been connected to the output of a component, the output impedance of the component, being vastly lower than the only impedance that was present in the setup (the input impedance of the scope), would have made it a completely different story.

Regards,
-- Al

Here's the stupid question: ... What would happen if you "simply [broke] the chassis to chassis shield connection" in an effort to eliminate interchassis ground current?

Hi Bryon,

That isn't a stupid question. The answer is that you would also be breaking the signal return path. Keep in mind that electricity only flows if there is a complete circuit, from source to destination and back.

If you were to break that return path in the interconnect cable, and both components had 3-prong power plugs, the return path that would be taken would be through the ac safety ground wiring of the power cords, and whatever additional safety ground wiring would be necessary to reach a ground point that is common to both components. That would result in noise problems.

If either or both of the components had a 2-prong power plug, and no other ground path existed in the system between the components, then the destination component would see a "floating" input voltage (i.e., one that is not referenced to its own ground). That would result in uncontrolled behavior, and very conceivably in full-power noise levels that could damage the speakers.

Best regards,
-- Al

I tend to think of "openness" as correlating with upper treble extension. For a line level interface, what would maximize upper treble extension (or at least minimize any degradation of it) would be low cable capacitance (meaning low capacitance per unit length and/or minimal length), and also low output impedance of the component that drives the cable.

Just a thought to keep in mind.

Best regards,
-- Al

Are there any generalizations about the typical effects of shielding on cable capacitance? Does shielding increase capacitance?

No, I don't think any such generalizations can be made, Bryon. Many shielded interconnects having low or very low capacitance are available from Cardas, Nordost, and Blue Jeans, among others. Unshielded interconnects with similarly low capacitance are available from Kimber and Nordost, among others.

The reason I cited those particular examples, btw, is that they come to mind as manufacturers who provide complete specs. It's unfortunate that a lot of other manufacturers don't do that.

I should add, btw, that my statements about the importance of capacitance pertain only to interconnects, not to speaker cables, where it is generally an unimportant parameter. Inductance can be important in a speaker cable. If it is high it can sometimes limit upper treble extension to a perceptible degree, especially with speakers whose impedance at high frequencies is low (electrostatics being notable examples).

Best regards,
-- Al