Unbalanced cable "directionality"... mismatched instructions?!?!

[lgo51]

I have read...

In Benchmark Media Systems' A Clean Audio Installation Guide it is said; "Edgar Lee Howard(11) has shown that a technical advantage exists for tying the shield of an interconnect cable at the send end rather than at the receive end. The potential advantage is a reduction in coupling of high frequency noise to the audio signal due to 'line to shield' capacitive unbalances."

And, I have read...

In Ground Loops Kent English of Pass Lab says; "Shield ground should not be connected on the source end of the wire, only at the input component end; label them and don’t forget!"

So, with all things considered, with a particular focus on equipment commonly found these days... Which is best?

Maybe one should connect the shield at both ends and sever it in the middle of the cable???????

Color me confused,
LarryO

[Steve Eddy]

Originally posted by lgo51
So, with all things considered, with a particular focus on equipment commonly found these days... Which is best?

Well, just make up a set of cables with the shield connected to just one end and try them both ways and see if one or the other way sounds better to you and if so, whichever way that happens to be, go with that one.

Maybe one should connect the shield at both ends and sever it in the middle of the cable???????

Hahaha! Good one!

Of course you may also consider dispensing with a separate shield entirely. Twisted pair, twisted/braided quad geometries are largely self-shielding to begin with.

Personally I think separate shielding is rather overrated for most situation.

se

[lgo51]

I think the same issue remains with briaded wires, only 2 of 3 carry signal information. Regardless of geometry, one is still functioning as a 'shield'.

Possibly I was not clear with my request. In the end, I'm sure we all select what pleases our ears most. But here, given two references from qualified sources, we find opposing recomendations, and hence a dilema. What I am seeking now is sound engineering, not engineering for sound.

Cheers,
LarryO

[Steve Eddy]

Originally posted by lgo51
I think the same issue remains with briaded wires, only 2 of 3 carry signal information. Regardless of geometry, one is still functioning as a 'shield'.

I don't know what sort or braiding you had in mind, but for the braiding I have in mind, all of the wires carry signal information.

Possibly I was not clear with my request. In the end, I'm sure we all select what pleases our ears most. But here, given two references from qualified sources, we find opposing recomendations, and hence a dilema. What I am seeking now is sound engineering, not engineering for sound.

Ok. If the issue is shielding, then the cable shield should be an extension of the chassis shielding and connected at both ends.

se

[Nelson Pass]

I am of the opinion that either way works, and for that
matter, if the (balanced) equipment is properly designed,
you can ground the shield to the chassis at both ends.

[lgo51]

Originally posted by Steve Eddy
I don't know what sort or braiding you had in mind, but for the braiding I have in mind, all of the wires carry signal information.

3 wires braided together...
Signal+ to Signal+
Signal- to Signal-
N/C to Signal- (connector shield)

I think that the 3rd, 'shield', wire only carries induced EMf/RFI currents. If i understand correctly the idea is to use the common mode rejection capability of the Target to cancel the induced noise on both the high impedance 'shield' and the high impedance Signal+ lines. Yes? No?

Cheers,
LarryO

[lgo51]

Originally posted by Nelson Pass
I am of the opinion that either way works, and for that
matter, if the (balanced) equipment is properly designed,
you can ground the shield to the chassis at both ends.

Thank you, Nelson.

All balanced would be heaven, but i'm wrestling with many unbalanced interconnects and just trying to get the odds going in my favor - and have a 'system' to rely upon as things get shuffled around, as they always seem to do.

Now if i could only get all the manufacturers to float the signal connectors off chassis ground and keep the mains safety ground away from the signal path entirely... and maybe win the lottery while i'm at it

Highest regards,
LarryO

[Steve Eddy]

Originally posted by lgo51
3 wires braided together...
Signal+ to Signal+
Signal- to Signal-
N/C to Signal- (connector shield)

Typically in a 3 wire braid (which I've always thought rather silly except perhaps for balanced lines) the thrid wire is doubled up with the ground lead.

I was speaking of twisted/braided quads.

I think that the 3rd, 'shield', wire only carries induced EMf/RFI currents.

Carries them to where? It's not going to carry any current if it's not connected to anything at the other end. Gotta have a closed loop to carry current.

If i understand correctly the idea is to use the common mode rejection capability of the Target to cancel the induced noise on both the high impedance 'shield' and the high impedance Signal+ lines. Yes? No?

Which idea? The three wire braid idea? If so, I don't know what the idea is behind it with regad to unbalanced lines.

se

[lgo51]

Originally posted by Steve Eddy
Carries them to where? It's not going to carry any current if it's not connected to anything at the other end. Gotta have a closed loop to carry current.

A current will flow in a conductor that experiences an [electro]magnetic field. Consider your radio or TV antenna, no physical closed loop there but it does have currents determined by the wavelength(s) of the impinging EMF. They are induced in the conductor and flow to "ground". No different than what we're speaking about in regard to interconnect wires. The higher the impedance, the higher the voltage developed by the induced currents... that's the law.

Cheers,
LarryO

[moamps]

Hi,

I think that interconnects should be connected to the ground at the power amplifier side. The power amplifier is where the biggest transformation of energy happens and where all kinds of noise are likely to be generated. For that reason, the output amplifier's audio ground must be connected to the chassis ground and mains ground/earth. Sources like CD, preamps, etc. should have the chassis ground normally connected to the mains ground/earth whereas the signal-audio ground should be connected only to the amplifier signal ground. This principle allows us to avoid ground loops thru mains power connectors and remove the current which is induced by the noise in the shield internally (amplifier parts) and externally (all sorts of EMIs ) via the shortest possible path.

However, the level of EMI is reasonably low in a home environment so shielding interconnects isn't a must.

Originally posted by Nelson Pass
.... if the (balanced) equipment is properly designed,
you can ground the shield to the chassis at both ends....

In professional systems (long multicore cable run), the shield is rarely connected to both sides (for example, between two radio control rooms and a master control room, where I always prefer connecting the shield at the master control room side). The only exception are cables for condenser mikes, which carry phantom power to mikes.

Also, a distinction should be made between unbalanced (usually using chinch connectors) and balanced connections (XLRs).
In unbalanced connections, the cinch connector's ground always represents the audio-signal ground and must be connected on both sides (with the cable's shield or second hot wire) because the ground carries the audio signal.
In balanced connections, XLR's ground pin (1) may be connected to the chassis ground and signal ground or not connected at all because the audio signal doesn't flow thru this wire (in correctly designed systems).

Regards,
Milan