Desirable electrical qualities for line level interconnects

[babysnake]

Hi

Regardless of how they are achieved, from an electrical and electronic engineering perspective, are there a series of agreed optimum electrical properties for an interconnect to possess?

I'm thinking resistance, inductance and capacitance values primarily

Cheers

 

[RayDunzl]

Just guessing, the optimum for an analog cable would likely be 0 Ohms, 0 Henrys, and 0 Farads.

Can't do that, so here's a couple of engineered cables suitable for interconnects;

http://www.canare.com/ProductItemDisplay.aspx?productItemID=53

https://catalog.belden.com/index.cfm?event=pd&p=PF_1192A

Scroll down for the specs.

 

[Ron Texas]

$1,000 per foot, LOL

 

[Wombat]

Tiresome topic.

Let's have a 'sticky' a 'beginner's tech.' thread and refer all inquiries to there.

 

[amirm]

This is the beginner forum so let's let the questions come in.....

 

[solderdude]

Funny plot for the Canare corp that shows the frequency response to be flat to 20kHz and drops 30dB at 200kHz already.
That's more than the 20dB/decade one would expect.
Perhaps an L-C-R thing ?
What source resistance and load resistance was used ?
Why the sharp 'knee' at 20kHz ?
Why use 100m ? if you want to impress someone with a plot use 1m

All the interconnects I ever measured had no problems from DC to 1MHz and have never run into problems with the usual 100pF to 200pF/m.
Resistance is a non issue for interconnects as source resistances are always relatively low and load resistances relatively high opposite each other.
Inductance also never has caused me any trouble in and well outside of the audible range.

The point where capacitance and inductance starts to matter is way beyond the audible range and of interest for video, RF or digital connections alone.

 

[Wombat]

This is the beginner forum so let's let the questions come in.....

Another 'senior moment'. This is getting embarrassing. At least I can still spell 'embarrassing'.

 

[SIY]

These are about the absolute least critical wires in the whole signal chain. Assuming a decent source and a non-pathological load, it has been demonstrated that copper, silver, and a potato can't be distinguished blind if levels are matched.

Worry about decent connectors, shielding, and construction. If it costs more than $10 for a meter, you're being cheated and may be getting an inferior product.

 

[babysnake]

These are about the absolute least critical wires in the whole signal chain. Assuming a decent source and a non-pathological load, it has been demonstrated that copper, silver, and a potato can't be distinguished blind if levels are matched.

Ok accepted

But what are a desirable/acceptable range of L-R-C values for an an interconnect in a typical domestic environment (up to 2m length say)?

There must be values that would be considered incompatible with the interconnect performing its task (AC current passage, 20Hz-20kHz at least, no filtering based on frequency)

Outside of those, does it matter?

 

[SIY]

If your source can't handle any non-pathological wire, replace your source.

A quick voltage divider calculation will convince you that even if the wire has stupidly high resistance (like two orders of magnitude worse that a typical 2m length of copper wire), it makes no real difference. And you'd have to really work hard to make a cable with enough capacitance over 2m to upset any competent source.

 

[babysnake]

If your source can't handle any non-pathological wire, replace your source.

A quick voltage divider calculation will convince you that even if the wire has stupidly high resistance (like two orders of magnitude worse that a typical 2m length of copper wire), it makes no real difference. And you'd have to really work hard to make a cable with enough capacitance over 2m to upset any competent source.

So the bar is set pretty low?

More or less any sensible electrical properties, decent connectors and shielding, copper will do

Anything over and above is unjustifiable, electrically speaking?

 

[SIY]

You got it. Avoid anything that boasts "hand-built" or special materials or whatever. Decent connectors are probably top priority, followed by shielding. These will work perfectly well, for example.

Michael Mardis's infamous blind experiment comparing copper to potato to banana to mud was pretty entertaining, especially the aftermath when people were post hoc rationalizing why their choices were random. And of course, once the choices were revealed, the attacks on the test validity were amusingly predictable.

 

[babysnake]

That makes, for all practical purposes, for an entire 3rd party cable industry based on technically correct but practically irrelevant science?

 

[SIY]

Often not even technically correct. And yes, when technically correct, practically irrelevant.

There may be an honest "high end" cable vendor out there, but in the thirty+ years I've been around them, I haven't met one yet.

 

[RayDunzl]

That makes, for all practical purposes, for an entire 3rd party cable industry based on technically correct but practically irrelevant science?

Wanna spend $100k or so on cables?

https://www.masterbuiltaudio.com/ultra-line-audio-cables

They may not be any good since they don't have magnets.

 

[Rod]

With 2 foot xlr interconnects for my dac/amp setup I thought just about any cheap cable would work just as good as the expensive ones because of such a short run. Guess I was right. Ended up with these: https://www.amazon.com/gp/product/B017C1D2R2/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

 

[solderdude]

So the bar is set pretty low?

More or less any sensible electrical properties, decent connectors and shielding, copper will do

Anything over and above is unjustifiable, electrically speaking?

Exactly...

To illustrate: a cable with 1nF capacitance (say 5 to 10 meters shielded wire) is connected to a source of 1kOhm (which is much higher than usual) will give a -3dB roll-off at 160kHz (-0.5dB around 50kHz)

A normal say 1.5 meter interconnect (say 200pF) connected to a CDP or DAC for instance with a 100 Ohm output R will give 8MHz as a -3dB point.
When the output R is 100 Ohm or lower even 10nF is not really problematic with a 160kHz cut-off point.
So cable capacitances even up to 1nF are not really an issue for interlinks.

An interconnect with a VERY poor resistance could be 10 Ohm (is kinda extreme). When your source is 100 Ohm and your load is 10kOhm the difference between a 10 Ohm and 0 Ohm cable would be smaller than 0.1dB.
Resistance is a non issue for interlinks.

Cable inductance: Say 1uH/m so a 10m cable is 10uH. The source resistance is 10kOhm (unusually high so worst case) = 160MHz
Non issue for audio at all.

 

[babysnake]

Deeply grateful for the explanation

Less impoverished too

So now it's just down to sleeve colour

 

[Djsedaw]

 

[DonH56]

^^^ You do know that string is not a conductor, yes?