advertisement


Interconnects - shielded or unshielded?

This topic reminded me of something I was reading recently about common-mode noise and interconnects - this seems very relevant:

A problem occurs when there is a ground voltage (common-mode voltage) between the two interconnected devices. Because of this voltage, a small current will flow down the cable shield between the devices (often referred to as common-mode current, or as a ground loop current). If the cable shield were ideal (zero impedance) this current would not cause a problem. However, since the shield has a finite resistance, a small noise voltage will appear across the length of the cable shield. The magnitude of this voltage will equal the common-mode current times the shield resistance. This voltage is in series RCA Phono Plug. Being an unbalanced system it is susceptible to common-mode noise voltages with the signal voltage and will add directly to it at the receiver. In other words, an unbalanced interconnect system consisting of only two conductors (center conductor plus a shield) has no ability to reject common-mode noise voltages).
 
Jim will be the authority here, however, I suspect that Kimber three-wire braided interconnects are a take on twin-lead balanced transmission lines, but for using the third wire as spacer, rather than encase the twin leads equidistant apart within a flat polyethylene ribbon (and then twist that along its length to assist with noise rejection from nearby sources).

With RF signals the uniform spacing of the wires is the key to the cable's function as a transmission line, as any abrupt changes in spacing would reflect some of the signal back toward the source, otherwise differences are cancelled. Normally this isn't a problem at audio frequencies and, especially, at line levels, however, Ray Kimber's early experiences installing Disco sound and lighting systems led to him using quasi-Litz type twisted speaker cables rather than steel conduit for the extreme long speaker cable runs that were otherwise negatively affected by strobe lighting electrical runs.

Basing a full line of home audio tonearm cables, interconnects, and speaker cables on this finding seems a bit of a stretch.
 
Tony L said:
The Kimber PBJ (and I think most Kimber stuff) is interesting in that it is a three core open weave with one core described as a ‘drain wire’ and only connected at the source end. I assume this is their take on the source-end screen, though I’ve never understood if this is just marketing or if there is actually any engineering credibility to it. I don’t understand how part of an exposed weave could have any screening properties.
Click to expand...

You need three wires (or more) to be able to make an alternating 'weave' as distinct from a simple twist! The only 'technical papers' I've seen on this from RA / Ben Duncan were, erm, unimpressive. [Ahem! I have used gentle phrasing here...].

An unsupported twist risks unwinding. A weave can avoid that.
 
Craig B said:
Jim will be the authority here, however, I suspect that Kimber three-wire braided interconnects are a take on twin-lead balanced transmission lines, but for using the third wire as spacer, rather than encase the twin leads equidistant apart within a flat polyethylene ribbon (and then twist that along its length to assist with noise rejection from nearby sources).

With RF signals the uniform spacing of the wires is the key to the cable's function as a transmission line, as any abrupt changes in spacing would reflect some of the signal back toward the source, otherwise differences are cancelled. Normally this isn't a problem at audio frequencies and, especially, at line levels, however, Ray Kimber's early experiences installing Disco sound and lighting systems led to him using quasi-Litz type twisted speaker cables rather than steel conduit for the extreme long speaker cable runs that were otherwise negatively affected by strobe lighting electrical runs.

Basing a full line of home audio tonearm cables, interconnects, and speaker cables on this finding seems a bit of a stretch.
Click to expand...

The behaviour depends on if the 'third wire' is simply a dielectric spaced or contains metal. In general twisting or braiding two or more wires tends to increase the capacitance per meter and maybe also the inductance per meter. In effect, it makes the cable electrically 'longer' by slowing the EM propagating down.

When I experimented with three wire weaves I never made a long enough run to accurately measure it. Too tedious to make by hand, you really need a mechanical setup to weave a long run.

However as with twin feed, it will tend to be far more sensitive to its surroundings than a sensible coax.

Note that 'Litz' is different because althought each strand has to be insulated from the others, they are all used in parallel as one 'wire'. This is to reduce 'skin effect' (more correctly called internal impedance or something similar).
 
I just use ferrite clamps. Bought a bunch many years ago in all sort of sizes and they just have a habit of finding their way around cables.

No idea if it makes any difference but I like the look of the little blighters.
 
YNWOAN said:
So, how about this, if one is using an interconnect that uses a separate shield (so not a coaxial construction) and that shield is protecting the internal signal cables from RFI, isn’t that picked up noise being connected to the signal ground? I mean, the vast majority of these types of interconnect (often called semi or pseudo balanced) do connect the shield to the signal ground.
Click to expand...
That may be possible if the screen acts as an antenna and there is enough EMI around. I ought to go back and sketch some example circuits and see. It's one of the basic problems of two-conductor-only interconnects for signal and screen that there are many ways in which the signal reference points and screens are related, and that can make a difference to noise floor. Which reminds me why I like the studio approach.
 
drummerman said:
I just use ferrite clamps. Bought a bunch many years ago in all sort of sizes and they just have a habit of finding their way around cables.

No idea if it makes any difference but I like the look of the little blighters.
Click to expand...

Beware of ferrites on speaker leads. Especially avoid placing them on just one conductor as they will easily saturate and cause bad distortion.
 
i'd been mostly kind of agnostic on the cable front - picking up good deals that were known to perform - but I'll be damned if my newer Vovox (unshielded) microphone cable doesn't sound LEAGUES better than my various other mogamis et.al. - they are a swiss company deep into the 'unshielded' philosophy and seem to be doing very well in audio - i have some vovox RCAs i also use in my phono circuit which are maybe a bit more detailed or maybe just kind of 'peaky' in the high frequencies - hard to tell - but the mic cable is surprising to say the least
 
Since the introduction of digital back in the 1980s and the 2v output "industry standard" for line level devices since then, it has been quite difficult to find good screened audiophile cable (particularly in silver) as most HiFi cable manufacturers dumped the screen as it was seen as not needed due to the higher line level voltages now being used. Also it has been very well documented over the years that unscreened cables sound superior to screened cables on live level signals and above (i.e. pre to power etc). I am referring to single ended connections here and not balanced. However, if you are using a vinyl front end then a screened tonearm cable either from the tonearm to the phono stage or from the tonearm to SUT and SUT to phono stage is mandatory.
 
Earth loop current flowing on a coaxial cable shield does not automatically appear as a large noise signal voltage. This coupling is the transfer impedance. A double shielded coax is much better.
 
Arkless Electronics said:
Beware of ferrites on speaker leads. Especially avoid placing them on just one conductor as they will easily saturate and cause bad distortion.
Click to expand...

Would you agree that power leads are the best place to use them? I see them on USB cables sometimes (maybe due to noise from the host device), but never really on analogue cables...
 
I guess it's all open to experimentation, but my understanding is that it would be a good idea to try them on power leads and other cables connected to equipment that you suspect is emitting noise. E.g. things like the power cables of your refrigerator, or HDMI cables from your television and things like this.
 
Arkless Electronics said:
Beware of ferrites on speaker leads. Especially avoid placing them on just one conductor as they will easily saturate and cause bad distortion.
Click to expand...

I'd emphasis the second point above.

It is usually OK to fit a clamp over a pair of wires (or a co-ax) as the 'out' and 'return' currents from the signal source then both pass together though the clamp. As a result the combined field in the ferrite will be low. However external *interference* picked up on the leads will usually be 'common mode' and the ferrite clamp preferentially affects it. There are some pretty diagrams illustrating this here:

http://www.audiomisc.co.uk/mains/filters2.html

Peronally, I'm happy enough with ferrite clamps on speaker cables provided you choose a large enough clamp and use it correctly. They shouldn't normally be needed, though.
 
davidsrsb said:
Earth loop current flowing on a coaxial cable shield does not automatically appear as a large noise signal voltage. This coupling is the transfer impedance. A double shielded coax is much better.
Click to expand...
Is that because the different shields will be optimised to have lower transfer impedance at different frequencies, ensuring a lower transfer impedance across the whole range?
 
adamdea said:
Is that because the different shields will be optimised to have lower transfer impedance at different frequencies, ensuring a lower transfer impedance across the whole range?
Click to expand...
come to think of it, that isn't really really relevant to earth loops is it? it will be the impedance of the earth loop (mains frequency?) which will matter I assume. Jim?
 
davidsrsb said:
A coax cable has a transfer impedance of about 0.01Ohm/m at audio frequencies, so if you have 1A !!! earth loop current in a 1m jumper, you should get 10mV noise. Read this article:
https://interferencetechnology.com/...shielding-factor-based-on-transfer-impedance/
Click to expand...
Yes. That's about what I was thinking and thanks for the article which is interesting.
I was just considering the point you raised about double shielding which is AFAIK normally applied to ensure that shielding is effective against different sorts of ambient interference (rather than to improve the noise position vis a vis ground loops).
As the man says, at very low frequencies transfer impedence is dominated by shield resistance, and the noise caused by a sane earth loop current in a well specified shield is not going to be huge.
 
adamdea said:
come to think of it, that isn't really really relevant to earth loops is it? it will be the impedance of the earth loop (mains frequency?) which will matter I assume. Jim?
Click to expand...

Short answer is the usual one: It depends. 8-]

Longer answer: *IF* the shield is fairly thick and has a generous cross-section of metal it may well will have a lower resistance/metre than the inner conductor. This then reduces the *cable* 'error voltage' caused by a ground loop current through the screen.

For RF in general you don't need a thick screen as the 'skin depth' is booger all. But for *audio* a thick outer can help keep externally induced currents on the *outer* surface of the coax, and this help reduce any effect in terms of interference injection or signal being affected by things up against the outside.

So if in doubt, a nice thick woven outer is attractive.

That's the theory. However in mere *reality* I've used long runs of 'sparse weave' UHF coax for decades without noticing any problems here. I bought a 100m drum of the stuff from RS back in the 80s and used it for TV/VHF/audio cables all over the place. (Tided parallel runs together with masking tape for stereo.)

Since then I tended to buy one of the 'Shark' brand coaxes that CPC sold. One that has a load of copper weave for the outer. Mainly because it fits nicely into the collet grips of the photo plugs I tended to use. Makes for a good secure bond between cable and plugs.

BTW in the system I used as my 'main' one for decades there is about a 5 meter run of coax to and from a QUAD 34 beside the listening chair. This uses the above cables and a 2-core (red, meant for gardening) mains cable all run side by side. Can't say it gave me any problems, and measured OK. Luck? Dunno.

(More specifically, 5m from the 730 preamp to the 34. 8m from the 34 back and out to the 732 power amp. All grounded back to the 730.)

FWIW we live just a few miles from what was a main RAF base, so loads of RF. They used the dome in our garden as a waypoint on their radars as they came in from the North Sea. Domes give a reliable reflection for radar. :) Although the RAF left the base, they still 'visit' on occasion.
 
You must log in or register to reply here.


advertisement


Back
Top