[warnerwh]

I am gettng some cable that is two twisted 20 guage wires, silver plated with teflon insulation and sheathing. This stuff has a braided silver plated copper shield.

My question is this: Do I connect the braided shield to ground or just the wire? Is there any theoretical improvement either way. I've seen some cables with the shield connected at one end and not the other. Being as I know diddley any help is appreciated.

Thank you

[SET Man]

Hey!

    What you could do is connect/solder the shield to the connector only on one side.... than only connect this side of the IC to the output of whatever component is going to be connect.

     This is why some or many I should say... IC have direction, notice some with arrow of which it should be connected.

    Hope this help.

Take care,
Buddy

[Ferdi]

Hi, probably too much detail....

Take good care that the shield is connected (with the neutral/ground wire) at one side only and the shield is well-insulated at the other end.

Ferdi

[Dan Banquer]

I'll disagree with the two posts above and for good reason. The general idea behind any good interconnect for unbalanced application is to connect both sides of the shield to ground. The reason for this is to minimize the ground impedance between the two units in question. This should give less noise, & better immunity from RFI/EMI. However given that there are no clear standards in grounding for consumer audio  and the propensity for so called audio designers to misunderstand grounding techniques you just might find that having the shield connected at one end does work out better.
Remember this: The cable you are making connects to two units, and the grounding scheme or lack of as the case may be will play a major part in your cable construction.
                d.b.

[Bill A]

The reason for this is to minimize the ground impedance between the two units in question.

Hi Dan,

Is the ground impedance minimized because of the equivalent larger gauge wire due to the shield and return side of the IC wired in parallel when both ends of the IC have the shield connected?  Is there a propensity for a ground loop within the IC when wired this way?

Bill

[Dan Banquer]

"Is the ground impedance minimized because of the equivalent larger gauge wire due to the shield and return side of the IC wired in parallel when both ends of the IC have the shield connected?"

Yes, and the braid will stay low impedance at higher frequencies.

"Is there a propensity for a ground loop within the IC when wired this way?"

No. Ground loops are caused by the equipment.
              d.b.

[andyr]

The general idea behind any good interconnect for unbalanced application is to connect both sides of the shield to ground.

Don't agree, Dan.

If the "shield" is just that - a shield - and there are two wires underneath it ('hot' and 'return') then connecting the shield at both ends puts it in the signal path (together with the 'return' wire).  I only use solid-core wire in my system so it is unacceptable to me to have braid as part of the signal path.

Hence I only connect the shield at one end (source) so the braid is earthed and thus able to act as a shield but is not in the signal path.

The reason for this is to minimize the ground impedance between the two units in question. This should give less noise, & better immunity from RFI/EMI.

I agree with you here!   

Using a thicker guage wire for 'return' will lower any ground differential.

Regards,

Andy

[warnerwh]

Speaking of guage what guage is necessary for IC's? I'll be using 20 guage but it seems to me that the average cd player usually outputs about 2 volts and most power amps need less than 2 volts for maximum output.

Being as we're dealing with so little current I'd think 20 guage is way overkill in this application. Are there any drawbacks to using thicker wire than 20 guage? Would we be better off with say 30 guage theoretically?

I see you want to keep capacitance down, does it become more or less as the wire gets thicker all else being equal?

Making my own ICs for myself and a couple of other people is only something fun to do. Also I can customize lengths which is nice. In reality what I end up with should be fine and well within  the boundaries of what is necessary. I'm just curious. Thanks

[andyr]

Speaking of guage what guage is necessary for IC's? I'll be using 20 guage but it seems to me that the average cd player usually outputs about 2 volts and most power amps need less than 2 volts for maximum output.

Being as we're dealing with so little current I'd think 20 guage is way overkill in this application. Are there any drawbacks to using thicker wire than 20 guage? Would we be better off with say 30 guage theoretically?

I see you want to keep capacitance down, does it become more or less as the wire gets thicker all else being equal?

Making my own ICs for myself and a couple of other people is only something fun to do. Also I can customize lengths which is nice. In reality what I end up with should be fine and well within  the boundaries of what is necessary. I'm just curious. Thanks

Hi Warner,

Thickness of wire for ICs is a contentious topic!   

You are quite correct in saying you don't need a large cross-sectional area of copper bcoz you aren't carrying much in the way of current.  But for ICs, there are the concepts of "skin effect" with too-thick wire or loss of bass through using too-thin wires which you have to take into account ... or ignore, if you think they're crap!   

I myself used 20g Vampire CC wire for my main ICs - between preamp and active XOs.  But I only did this because they are very long (35') and I wanted to minimise earth potential differences ... see Dan B's post above.

IMO, 24g is the thickest you want to go for normal length ICs up to 6 or 7'.   However, I'm currently in a phase of using 3 x 30g wires for 'hot' and 24g for 'return' (rather than just 24g wires for each) ... the reason for this is that (unlike the actress) I believe "thinner is better".  Trouble is, you have to be able to stop the thin wires from moving - if they can, you lose bass (this I have experienced from a simple experiment    ).  So 3 x 30g wires twisted together provides sufficient mass and stiffness to stop them moving around when the signals go through them.  Yes, I know ... hard to believe!!   

And capacitance is not (or only indirectly) affected by wire thickness ... it's the physical construction of the cable which influences its capacitance.  And BTW, IMO R & L are irrelevant for ICs - they only matter for speaker cables and PCs.  But C is important for ICs (too much C can roll off your treble).

Regards,

Andy

[ctviggen]

Don't agree, Dan.

If the "shield" is just that - a shield - and there are two wires underneath it ('hot' and 'return') then connecting the shield at both ends puts it in the signal path (together with the 'return' wire).  I only use solid-core wire in my system so it is unacceptable to me to have braid as part of the signal path.

Hence I only connect the shield at one end (source) so the braid is earthed and thus able to act as a shield but is not in the signal path.

In my former life as an electrical enginerd, this is what I was taught:  if  you ground both ends of the shield, you'll cause a ground loop. Thus, you only ground one end.  I would think the signal end would be the best end to ground, but I have no data to suggest this.

[Dan Banquer]

I think this is a common misconception related to unbalanced configuration and this may well be due to the fact that in unbalanced consumer audio we have high input impedance's at line level. This gives well below 1mA of signal current so since the "shield" never carries much of any current there is no ground loop so to speak. The second issue is the fact that this is unbalanced and the ground connection is there by definition so when you ground the shield at both ends there cannot be a ground loop caused by the wire. The real advantage of grounding the shield at both ends for unbalanced is that you reduce the ground impedance between the two units in question.
Ground loops for consumer unbalanced are in the units, not the wire connecting them. If you need to break a ground loop between two units you have a few choices, fix the grounding inside the units or get a line level transformer.
                   d.b.

[ctviggen]

Hmmmm...As always, it's more confusing than it seems.

http://www.rane.com/note110.html

He recommends grounding both sides of the shield to the chassis. 

This person, however, recommends only grounding one side of the shield:

http://www.benchmarkmedia.com/caig/html/caig05.html#anchor593510

I can't really tell which has the better position.

[andyr]

Don't agree, Dan.

If the "shield" is just that - a shield - and there are two wires underneath it ('hot' and 'return') then connecting the shield at both ends puts it in the signal path (together with the 'return' wire).  I only use solid-core wire in my system so it is unacceptable to me to have braid as part of the signal path.

Hence I only connect the shield at one end (source) so the braid is earthed and thus able to act as a shield but is not in the signal path.

In my former life as an electrical enginerd, this is what I was taught:  if  you ground both ends of the shield, you'll cause a ground loop. Thus, you only ground one end.  I would think the signal end would be the best end to ground, but I have no data to suggest this.

Re. "which end to ground" ... not being an EE either, I have no "scientific" basis to back this up but people seem to recommend grounding the source end as this, relatively speaking, is the low noise end.  Except in the case of phono cables which need to be earthed at the destination end.

Regards,

Andy

[andyr]

In my former life as an electrical enginerd, this is what I was taught:  if  you ground both ends of the shield, you'll cause a ground loop. Thus, you only ground one end.  I would think the signal end would be the best end to ground, but I have no data to suggest this.

Re. "which end to ground" ... not being an EE either, I have no "scientific" basis to back this up but people seem to recommend grounding the source end as this, relatively speaking, is the low noise end.  Except in the case of phono cables which need to be earthed at the destination end.

Regards,

Andy

[Bill A]

The references given by ctviggen pretty much deal with balanced I/O's, although the Rane reference does diagram single ended configurations. Check out diagram numbers 23 and 24.  In 23, the shield is tied to chassis ground at the input, or source, device only. In diagram 24, the shield is tied to the signal return and the chassis ground at the output, or destination, end.   Unless the I/O's have their returns isolated from the chassis, 23 and 24 are the same, differing only in which end of the IC has the shield connected to the chassis.  In neither case is the shield tied to the return or chassis at both ends.

I would think that for home audio use, all single ended I/O cables should have only one end of their shields tied to the chassis, and at a common point; the preamp, for example, seems like a logical choice for this.  The other end of the shield would be left floating.  This would not negate the need for all equipment to be at the same ground potential, but that is done through the IC return wire.

Bill

[sts9fan]

ok I am confused. I have read that attaching the sheild to one end is the way to go. The article here on ground loops is one of them.

http://www.passlabs.com/articles.htm

While it was not written by Mr Pass himself I do trust he would back it up and I respect his designs quite a bit. Can you please in laymans terms point out where thier theory is incorrect?

Thanks

[andyr]

ok I am confused. I have read that attaching the sheild to one end is the way to go. The article here on ground loops is one of them.

http://www.passlabs.com/articles.htm

While it was not written by Mr Pass himself I do trust he would back it up and I respect his designs quite a bit. Can you please in laymans terms point out where thier theory is incorrect?

Thanks

I think all that happened is Kent made a simple mistake.

At the top of p2, he first says: "Shield ground should not be connected at the source end of the wire - only at the input component end."

I think he meant to say "destination end" because he then goes on to say: "This will of course mean that at the input compnent end of (the) cable, the shield and ground signal conductor will bond together."

Regards,

Andy

[Bill A]

Well, doing a little net research tells me that I am wrong about connecting the shield only at the destination (pre-amp) end  as I had suggested.  So it appears that, like andyr said, shield connection at the source end is the way to do it for directional cables.

Also, some suggest using only coax cable (shield connected at both ends, of course), rather than shielded pair due to increased capacitance caused by that configuration. http://www.bluejeanscable.co.uk/articles/balanced.htm

Tara Labs has cross-sectional pictures of their cables, some shield pair, some not.  They also give a discription of the sound you will get with each.

Am curious, though, as to why use a shielded pair then tie the shield to the return at both ends?  It seems to me that a coax cable would work just as well in that configuration.

Bill

[Dan Banquer]

The low frequency field rejection of a twisted pair is certainly applicable to line level audio.
                  d.b.

[Bill A]

Are you talking about rejection of EMI due to, say, AC mains, just by using a twisted pair within the shield? (I assume that we are using single ended I/O's)

I'll have to cogitate a bit on why that is.

Bill