advertisement


Interconnects - shielded or unshielded?

Even with non-coax, say if you're talking about twisted pair, shields are still important. If you're looking at the digital side of things, there's a reason why Category 7 ethernet cables include a shield as part of the spec, and why they're rated for transmission of 600 MHz signals over 100 meters, as opposed to 500 MHz for Category 6a unshielded cables. For any sort of mission-critical data center use with 10 Gbps ethernet, they almost always use shielded cables these days, at any length. While our audio signals are not extending into that frequency range (hopefully), that need and spec arose out of measurable data loss with unshielded cables. It can be random and unpredictable. The recording industry has been using shielded XLR cables for microphones for the better part of a century. So I don't see why you wouldn't for audio interconnects.

The Kimber issue is a multi-layered thing. In addition to not being surrounded by a shield (as I understand it, their drain wire is only connected at one end of the interconnect), the geometry of their cables can make them capacitive. When I bought my first Naim amp, I was coming from a system with Kimber 8TC speaker cable, and it sounded absolutely dreadful. I didn't understand, because the cable sounded wonderful on all previous amps I had used it with. But that braid made it capacitive, which was an undesirable electrical quality in the context of Naim design. I switched it for some Chord Rumour-2, and immediate recovered the qualities that Naim is known for. I would imagine that similarly, Kimber interconnects are not a synergistic match for all systems, as much as I enjoy their sonic signature when they are a good match. PBJ was one of the great audio deals of the late-'90s, and I had their Silver Streak on my wish list until going Naim.
 
It may not be 'tied' to ground at one end, but *provide* the ground reference there.

Consider the traditional approach of only having the preamp connected to mains 'earth' and using the coaxes from it to the sources and power amp as their ground reference. i.e. they *aren't* themselves grounded to the mains earth so far as the signal is concerned, but use the coax outer as their 'ground' point. In effect, the outer of the coax makes them all behave as if they're all in one big metal box that is earthed.
Thanks- that’s helpful.
 
In a standard coax, yes, but not in cables like say vdH D102 which is a twin core cable with an outer coax screen that is connected at one end only.
Yes. I’m guessing that the VDH is not really coax and is using the second core as a return ie it’s a shielded (twisted?) pair. As I understand it (ie not very much) the fact that the screen is coaxial (has the centre of its cross section coinciding with the signal core) and takes the return current, is a crucial part of the electrical properties of coax which makes it not radiate rf.
If I read the blurb right the VDH app is a balanced cable with RCAs on either end.
 
Reading the above prompts me to point out that - once the coax outer is 'thick enough' - the signal ground return current will flow on the *inner parts* of the coax sheild. Thos this ensures better circular symmetry than having a twisted pair at the center. The twisted pair may be 'current' balanced. But if the signal is single ended, the signal on the pair won't be 'voltage' balanced.

Similarly, if the coax isn't connected at one end, then at that end its potential may vary, and will be capacitatively coupled to the inner wires.
 
Yes. I’m guessing that the VDH is not really coax and is using the second core as a return ie it’s a shielded (twisted?) pair. As I understand it (ie not very much) the fact that the screen is coaxial (has the centre of its cross section coinciding with the signal core) and takes the return current, is a crucial part of the electrical properties of coax which makes it not radiate rf.
If I read the blurb right the VDH app is a balanced cable with RCAs on either end.
I think it's usually called pseudo-balanced.

The characteristic of a fully balanced connection that gets you the external noised rejection is that the impedance to signal ground is the same on both conductors. As @martin clark wrote earlier the impedance of the interference source matters. If the two conductors see the same impedance from the interferer and from the signal source the cancellation will be good.

So to achieve a balanced connection a two-conductor cable needs cooperation from the signal source and destination. I suspect this is not common with RCA connections. In particular the cold connection at the source end needs to have the same impedance as the output impedance of the hot connection. A resistor in series rather than directly connected, which I am not sure is common.
 
Even with a twisted pair cable, the RCA connector is very unsuitable for some form of balanced circuit.
The ideal case would be a shield connected to case at both ends and the audio on balanced pairs into a differential input, from an impedance balanced source circuit.
A Cat-7 Ethernet cable would actually work very well, stereo in a single cable.
Too cheap for audiophile nervosa though
 
Do not have wires in a conventional sense in my interconnects made from 100% pure silver foil shielded and far superior to Kimber.
 
Sorry if I have got this wrong but isn’t the shield tied to ground at both ends in coax? In the case of a twisted pair it’s different.

Sorry for late response - yes it is, for a coax.

So for coax used in a single-ended connection (audio, using RCA connectors or sim) while the coax outer will shield the inner, the best/only real defence is to ensure that coax outer contains plenty of copper and dense coverage resulting in a very low impedance end-to-end, so that any induced currents in it from extraneous sources amount to negligible voltages end-to-end compared with the wanted signal (which interference voltage will appear in series with, for a single-ended interconnect). This is not particularly difficult for cables carrying audio over moderate distances in a home environment, in my view.

(I should have made clear in my previous post, most of it - unless explicit otherwise - was about single-ended interconnections; which really can be 'good enough'.)
 
My testing has been from pre to power . The hum increased when using unbalanced (Naim pre out ) to std RCA input on a power amp (tbh -- I suspect the grounding within the amp was the culprit). Tried a variety of additional shields (outer nor connected either end ) around a std RCA coaxial cable set . to no effect . Ended up using single ended out into balanced input (pseudo ?) with pin 2 hot for positive , pin 3 neg and pin 1 not connected to screen on the amp end . Turned out to be the quietest by far . No real logic , just trial and error . In my set up having the screen was a real help , but fully understand the risk of this being an antenna for others ................
 


advertisement


Back
Top