Ethernet cable for balanced audio?

[Old_School_Brad]

How would you suggest doing it then?

Connecting both wires of a twisted pair to the same source (e.g., both to the positive) would defeat the purpose of using a twisted pair. For noise rejection to work effectively, one conductor should carry the positive (or original) signal while the other carries the inverted (negative) version.

If both conductors in the twisted pair carry the same signal, the setup loses its differential advantage, and any external noise picked up by both conductors would not cancel out, thus reducing the noise suppression benefit. Instead, it would just act as a single-ended connection with higher susceptibility to interference.

 

[ppataki]

Thank you @unpluggged and @Old_School_Brad
Now I understand - I will rewire the cables accordingly

Just one more question for my benefit: there are cables like the Sommer Epilogue that have two twisted pairs (plus the shield) and the balanced connection is made by using one pair for hot and the other pair for cold. So is that wrong altogether and shall not be done?

 

[unpluggged]

So is that wrong altogether and shall not be done?

It is wrong in the sense that the common-mode rejection will be compromised, but the signal can still be transferred. In any case, this is not the way balanced cable should be constructed.

See here: https://en.wikipedia.org/wiki/Differential_signalling

 

[ppataki]

Thanks, Gents, much appreciated
I have been here for 5 years now and I learn new things almost every day

If I may have some more questions:
- if I wanted to connect four balanced connectors (let's say 6.3mm balanced jacks) using just one ethernet cable then I shall use one pair for hot-cold for each jack connector (i.e. that will consume all four pairs) and then what shall happen to the shield? Would the one shield be 'shared' with all four connectors? Not sure how to make it happen in practice
I am asking since with this I would just need one single cable for all my 3 channels....

and lastly:
- if this was an unbalanced connection, would all this still matter? Could I just - only for the sake of an example - twist all the four pairs together for the signal and connect the shield to the sleeve?

Thank you

 

[sergeauckland]

Thanks, Gents, much appreciated
I have been here for 5 years now and I learn new things almost every day

If I may have some more questions:
- if I wanted to connect four balanced connectors (let's say 6.3mm balanced jacks) using just one ethernet cable then I shall use one pair for hot-cold for each jack connector (i.e. that will consume all four pairs) and then what shall happen to the shield? Would the one shield be 'shared' with all four connectors? Not sure how to make it happen in practice
I am asking since with this I would just need one single cable for all my 3 channels....

and lastly:
- if this was an unbalanced connection, would all this still matter? Could I just - only for the sake of an example - twist all the four pairs together for the signal and connect the shield to the sleeve?

Thank you

The one shield should be shared over the four connectors by soldering four short tails to the shield wire and twisting the shield wire with the two signal conductors. You can cover the wires in heat-shrink or sleeving tubing for a neater appearance.

If you want to use that cable for a single unbalanced connection, then yes, twist the 8 wires into one, and use the shield to the sleeve. Somewhat wasteful of the cable, and hard to make it neat with 8 wires twisted together, but can be done. Bear in mind that the shielding in ethernet cable may not be as good as a 'proper' piece of coax, but at line level should be good enough. Better would be to use a single piece of RG6 or RG59 Coax!

S.

 

[somebodyelse]

Thank you @unpluggged and @Old_School_Brad
Now I understand - I will rewire the cables accordingly

Just one more question for my benefit: there are cables like the Sommer Epilogue that have two twisted pairs (plus the shield) and the balanced connection is made by using one pair for hot and the other pair for cold. So is that wrong altogether and shall not be done?

Firefox's internal translation isn't good for that page, but I think they're talking about a star-quad not two twisted pairs. That's a cable construction with better interference rejection than twisted pair. It's usually only used for low level signals like microphones in electrically noisy environments.
https://www.belden.com/blogs/broadcast/how-starquad-works
https://benchmarkmedia.com/blogs/ap...-the-importance-of-star-quad-microphone-cable

 

[ppataki]

Firefox's internal translation isn't good for that page, but I think they're talking about a star-quad not two twisted pairs. That's a cable construction with better interference rejection than twisted pair. It's usually only used for low level signals like microphones in electrically noisy environments.
https://www.belden.com/blogs/broadcast/how-starquad-works
https://benchmarkmedia.com/blogs/ap...-the-importance-of-star-quad-microphone-cable

In the meantime I have reached out to their support desk and they have confirmed that it was Star Quad design indeed, not twisted pairs

 

[Speedskater]

While there is no need to do it. In a CAT cable you can make the equivalent of a StarQuad balanced cable. Connect two solid colored wires to one terminal and their striped wires to the other terminal.

 

[Punter]

I worked in a radio station once that was completely wired up with telecom style, CAT-3 multi-pair cable. Every circuit was a twisted pair with no shield connection as the cable itself had no shield or drain wire. The only signals not run in telecom cable (up to 100 pairs) were the microphones which were short runs inside the studios. The whole station was noise and hum free. All of the balanced signals in the telecom cables were a similar level so there was also no crosstalk. There would have been runs of over 50 meters in the station, magical audio cables were not needed to maintain signal integrity.

 

[ppataki]

While there is no need to do it. In a CAT cable you can make the equivalent of a StarQuad balanced cable. Connect two solid colored wires to one terminal and their striped wires to the other terminal.

I tried to go for the StarQuad design
Please let me know if this is OK:

Thank you

 

[AnalogSteph]

Whatever this is, it certainly doesn't look like starquad to me.

If you have to use two pairs each, I would keep them twisted like they came out of the cable and use them in antiparallel, i.e.
Blue/white pair - blue to tip, white to ring
Green/white pair - white to tip, green to ring

The shield looks OK as-is, if not terribly shield-y... maybe leave it a bit longer and twist it around your two pairs each.

 

[ppataki]

Whatever this is, it certainly doesn't look like starquad to me.

If you have to use two pairs each, I would keep them twisted like they came out of the cable and use them in antiparallel, i.e.
Blue/white pair - blue to tip, white to ring
Green/white pair - white to tip, green to ring

The shield looks OK as-is, if not terribly shield-y... maybe leave it a bit longer and twist it around your two pairs each.

I followed @Speedskater 's advice above about how to wire it for StarQuad
Did I misunderstand your advice @Speedskater ?

 

[solderdude]

For balanced operation use a single pair for connection (so for instance blue on hot (tip) and blue-white on cold (ring).
Then use the adjacent pair (tied together) for the screen (sleeve).

When doing XLR use screened wire and use the drain wire for all screens (so not pin 1 on XLR).

 

[somebodyelse]

While there is no need to do it. In a CAT cable you can make the equivalent of a StarQuad balanced cable. Connect two solid colored wires to one terminal and their striped wires to the other terminal.

Please explain how that is equivalent to star-quad for interference rejection.

 

[Speedskater]

In a StarQuad the increased cancellation is with 2 Hots & 2 Colds their exposure to noise/interference is more symmetrical and balance is all about symmetry.
So in a CAT cable if we wire:
Blue to Hot and Blue/White to Cold
Red to the same Hot and Red/White to the same Cold
Because the two pairs have different twist rates, the likelihood of increased cancellation occurs.
But with CAT cable the improvement will be minuscule. However it will increase mechanical robustness.

The is no need for a screen/shield pair on a line level interconnect. Only needed for phantom powered mics.
* * * * * * * * * * * * * *
Also if the CAT cable has a shield/screen with a drain wire, the two pairs may reduce the possibility of Shield Current Induced Noise (SCIN).

 

[somebodyelse]

In a StarQuad the increased cancellation is with 2 Hots & 2 Colds their exposure to noise/interference is more symmetrical and balance is all about symmetry.
So in a CAT cable if we wire:
Blue to Hot and Blue/White to Cold
Red to the same Hot and Red/White to the same Cold
Because the two pairs have different twist rates, the likelihood of increased cancellation occurs.
But with CAT cable the improvement will be minuscule. However it will increase mechanical robustness.

It's the relative positioning of the hots and colds in the quad that is important for the interference rejection though, not just that there are 2 hots and 2 colds. I still don't see how using 2 of the pairs in the CAT will give equivalent rejection. Do you have any references explaining it?

 

[raindance]

It is better to use twisted pairs rather than doubling up. There's no significant current carried so no need to double up at all. Noise rejection will be better.

 

[Hayabusa]

Does anybody here have experience with using an ethernet cable as a balanced interconnect?

I was thinking that I would give it a try, just for fun
I was searching for it online and apparently I am not the first to think about this:
View attachment 402495

Unlike in the above picture I am planning to use 3 pairs of cables for 1 channel (so one pair will be unused)

I got this cable, it is Cat 8.2 with 4 pairs of AWG22 cables

View attachment 402496
View attachment 402497

During the weekend I shall have time to finish all the soldering and then I will give it a try in my home office system (where I currently use Sommer Epilogue interconnects)

I did test this, in my case I had to use more that one ground wire to get the hum down....

 

[JP]

In a StarQuad the increased cancellation is with 2 Hots & 2 Colds their exposure to noise/interference is more symmetrical and balance is all about symmetry.
So in a CAT cable if we wire:
Blue to Hot and Blue/White to Cold
Red to the same Hot and Red/White to the same Cold
Because the two pairs have different twist rates, the likelihood of increased cancellation occurs.

I don’t think so:

“Star quad cables often use filler elements to hold the conductor centers in a symmetric four-point arrangement about the cable axis. All points of the star must lie at equal distances from the center of the star. When opposite points are connected together, they act as if they are one conductor located at the center of the star. This configuration places the geometric center of each of the two legs of the balanced circuit in the center of the star. To a magnetic field, both legs of the balanced circuit appear to be in the exact center of the star. This means that both legs of the balanced circuit will receive exactly the same interference from the magnetic field and a common-mode interference signal will be produced.”

 

[rcstevensonaz]

To restate what others have said in slightly different words, increasing the AWG of the signal wires does not have any benefit for audio interconnect cables.