[Gingerbread Man]

Am I right in the assumption that it is good practice to connect a shield to only one end? Say I have a twisted pair, signal and ground, soldered onto RCA phono plugs, inside a shielded cable, so I would twist and solder the shield at the end onto only one plug?

Also, does it matter which end the shielded plug is plugged into (i.e. source or amp)?

 

[majkel]

I tried many combinations with two wire shielded cables and for unbalanced/RCA it is best to use just one wire for the signal, and the shield for the ground. You leave the second wire alone. This way you get the lowest capacitances from signal to ground.
However - if you want to try - you use one wire for the signal while the other for the ground, and you connect the shield to the plug at the source side.

 

[Gingerbread Man]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif I tried many combinations with two wire shielded cables and for unbalanced/RCA it is best to use just one wire for the signal, and the shield for the ground. You live the second wire alone. This way you get the lowest capacitances from signal to ground. However - if you want to try - you use one wire for the signal while the other for the ground, and you connect the shield to the plug at the source side.

Well I was going to use a coaxial cable, strip out the signal wire and dieletric, and thread a twisted pair of cat 5 through it, taking use of the foil and braiding as a shield whilst using the cat 5 as signal and ground - would this have more detremental benefits than good rather than just using the twisted pair?

 

[error401]

Quote:

Originally Posted by Gingerbread Man /img/forum/go_quote.gif Well I was going to use a coaxial cable, strip out the signal wire and dieletric, and thread a twisted pair of cat 5 through it, taking use of the foil and braiding as a shield whilst using the cat 5 as signal and ground - would this have more detremental benefits than good rather than just using the twisted pair?

It's probably not worth the trouble, I wouldn't think. You'd get similar characteristics from shielded twisted pair.

Check out this paper on grounding and shield connections. It's very interesting: Grounding and Shielding Audio Devices

Their bottom line is to connect the shield at both ends to chassis ground, but you probably will have to connect the shields to signal ground which isn't optimal. Either way they recommend grounding both shields.

 

[Thermionic]

A long time ago I think Mati Otala (sp?) suggested that for best SQ both the signal and the ground or return wires should be of similar construction and made out of the same material. As you probably know cheap screened audio leads might often have a tin plated outer braid and an un-plated inner conductor of say 7 strands of copper. This imbalanced structure of a few copper strands of the inner conductor to many more plated strands of the screen was thought less than ideal, and Tin plating was thought to degrade the sound as well. So some experimenters and manufacturers (Van Den Hull is one I know) started using similar ground and return wires screened by an overall braid connected at just one end to avoid hum loops. The convention was to have the screen connected at the amplifier end because that would (in the UK at least) have a earthed plug into the mains, but I have no idea what the latest safety standards require.

What you propose in my own limited experience seems Ok, in fact in my DIY days it was popular to use un-screened 300 Ohm ribbon between a pair of phono plugs, which worked well if you kept it away from sources of hum and I think DNM use a single stranded refinement of this even today. I also used to get air spaced TV down lead and pull out the single inner conductor, then I would thread a pair of silver wires in opposing voids to make an air spaced pseudo balanced cable which had the outer braid connected at one end only. I also thought of doing this with silver plated wire wrap wire or even plain solder sometime. I was interesting to experiment and sometimes things seemed to be an improvement, though at other times not.

 

[infinitesymphony]

I've read a lot lately about connecting the shield at only one end of a cable, including both balanced and unbalanced types. Now I have two questions:

1. Is this why some manufacturers have directional arrows on interconnects?
2. Why do most manufacturers connect the shield to both ends (nondirectional)?

 

[Negatron]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif I've read a lot lately about connecting the shield at only one end of a cable, including both balanced and unbalanced types. Now I have two questions:

1. Is this why some manufacturers have directional arrows on interconnects?
2. Why do most manufacturers connect the shield to both ends (nondirectional)?[/QUOTE]

EDIT: I hit save before typing, see below.

 

[Negatron]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif I've read a lot lately about connecting the shield at only one end of a cable, including both balanced and unbalanced types. Now I have two questions:

You will find an abundance of 'expert' advice on the Internet about connecting the shield of a twisted pair. It largely depends on the specific application and much of what you read is based on RF or Radio Frequency use, whether copied and re-written for audio or not.

Tight Twisted Pair has pretty good RFI & noise rejection in itself. From experience, plus test data over the years I suggest the best way for noise rejection in the AUDIO spectrum is to connect the twisted pair shield ONLY at the source end. That shunts all noise back to the source ground and not the input. It is more indirect.

Another thing besides noise, is that wire has a grain structure and orientation makes a small difference. Good luck figuring out which direction is best and maintaining your sanity.
However, keep the labeling the same direction so both channels match. We wouldn't be here if we weren't a little nuts, so it can't hurt. Right?
Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif 1. Is this why some manufacturers have directional arrows on interconnects?

They likely understand Audio signals and know what they are doing

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif 2. Why do most manufacturers connect the shield to both ends (nondirectional)?

Provided the cable is shielded twisted pair and not coax or twin-ax, draw your own conclusions based upon #1.
plus, It costs more in production time to properly shield only one end.
JMPO

 

[error401]

Quote:

Originally Posted by Negatron /img/forum/go_quote.gif 1. Is this why some manufacturers have directional arrows on interconnects? 2. Why do most manufacturers connect the shield to both ends (nondirectional)?

[/QUOTE]

Read the link I posted a few posts up. It's from a professional audio gear manufacturer and goes into the second issue in detail.

Short version: It's better to connect the shield at both ends.

The slightly longer story is that you should connect the shield to chassis ground at both ends, if possible. In essence, this creates a Faraday cage that encloses both chassis and the signal wires inside the shield. There's no way for EMI to penetrate anywhere.

Connecting the shield at only one end creates an antenna on that end's ground. If EMI is picked up, the ground at that end of the signal will be modulated, and therefore the signal will be modulated as well. If both ends are connected, any potential induced by EMI should be better balanced between the two grounds and should cause less distortion. The real reason this has become common practice is to avoid hum-inducing ground loops, which occur due to poor or improperly designed grounding.

Sure, TP has some rejection on it's own, but using the shield properly costs you nothing and it can't hurt

.

 

[Negatron]

Quote:

Originally Posted by error401 /img/forum/go_quote.gif Read the link I posted a few posts up. It's from a professional audio gear manufacturer and goes into the second issue in detail. Short version: It's better to connect the shield at both ends. The slightly longer story is that you should connect the shield to chassis ground at both ends, if possible. In essence, this creates a Faraday cage that encloses both chassis and the signal wires inside the shield. There's no way for EMI to penetrate anywhere. Connecting the shield at only one end creates an antenna on that end's ground. If EMI is picked up, the ground at that end of the signal will be modulated, and therefore the signal will be modulated as well. If both ends are connected, any potential induced by EMI should be better balanced between the two grounds and should cause less distortion. The real reason this has become common practice is to avoid hum-inducing ground loops, which occur due to poor or improperly designed grounding. Sure, TP has some rejection on it's own, but using the shield properly costs you nothing and it can't hurt .

I have read the link posted in the past as well as many others, as well as opinions of equal credential stating the benefit of gnd at source only. I still stand by my opinion. I am in agreement in shield from chassis to chassis only with TP if that were commonly available, but it is not. In most systems the gnd side of the signal is connected to pwr gnd, and at both ends. Thus the EMI and RFI is injected into both power supply negative, thus wired in paralell to the signal. and the following supply being the beginning of another amplification chain.

As I posted previously, I wrote as opinion only. I will add to that by saying that when I began building and modifying Audio gear there was a great abundance of scholarly works stating that interconnects had no effect on the sound, wire was wire! There was not a multi-million industry selling cables and a fair share of snake oil as there is today. I would also add that papers were published by the AES that concurred with the commonly held beliefs that wire was not audible. Oh BTW, in the late 1800's the National Academy of Sciences proved beyond a shadow of a doubt that it was physically impossible to exceed 100MPH.

I am rather odd, I actually voice my designs and use my ears. Not trusting test gear as being as sensitive as my hearing when it comes to things we do not know how to measure or quantify yet. But as I stated, it is Just My Opinion, and since the issue is not going to be resolved anytime soon by large numbers of 'Experts' I see no real purpose in debating the issue further. I prefer to leave that to the afore mentioned "Experts".

As for the original poster, build both with the same cable and termination and listen.

 

[infinitesymphony]

Quote:

Originally Posted by error401 /img/forum/go_quote.gif Read the link I posted a few posts up. It's from a professional audio gear manufacturer and goes into the second issue in detail. Short version: It's better to connect the shield at both ends. The slightly longer story is that you should connect the shield to chassis ground at both ends, if possible.

That's interesting... I did skim through that, but I have another document from Rane that has some conflicting imagery. It says that the "right way" to make a cable is to connect the shield at both ends using signal ground for pin 1, and the "recommended way" is connecting the shield at both ends using chassis ground for pin 1, but then it shows connection diagrams where the shields are only connected at one end.

Here's a link to the HTML version: Rane - Sound System Interconnection.

I asked the guys at Blue Jeans Cable yesterday, because they usually have a no-nonsense take on these matters, and here's what they said:

Quote:

"We don't follow this line of thinking as it doesn't solve the ground loop problem, but only acts as a 'band aid' for it. Furthermore, on some equipment, connecting the cable in the incorrect direction can cause some nasty damage to the electronics. Ground loops are best eliminated at the power supplies of your various gear. Sometimes having gear plugged in at different circuits in your home can cause ground loops. Trying to 'fix' the problem through cabling is usually not the best route."

 

[jcx]

with RCA-RCA audio interconnects you have single-ended signals which are referenced to the shield gnd

the best practical single-ended audio signal transfer comes from using heavy shield braid coax with very low shield(=gnd ref) DC-AF resistance and high symmetry coaxial construction/termination to give good EMI/RF shielding effectiveness

Quote:

Originally Posted by jcx /img/forum/go_quote.gif MOGAMI - The Cable of the PROS or if you can diy: "'Exotic cables will not stop noise. Expensive cables, even if double or triple shielded, made of 100% unobtainium, and hand woven by virgins will have no significant effect on hum or buzz. Only the resistance of the grounded conductor can make a difference.' If you have to use coax, Belden 8241F, with a low-resistance copper-braid shield, works well for audio and video." from: Washington June AES - SMPTE.org

twisted pair with balanced construction is better for balanced xmit/receiver connections, then an overall shield may be helpful depending on the electronics’ CMRR, and the EMI environment, the EMI and system consideration (ground loop noise) then can dictate optimum shield connections - hybrid ground systems are open on one end at DC, with good RF shorting at both ends, quality systems using special feedthru cap construction

 

[jcx]

.

 

[Seamaster]

Quote:

Originally Posted by error401 /img/forum/go_quote.gif It's probably not worth the trouble, I wouldn't think. You'd get similar characteristics from shielded twisted pair. Check out this paper on grounding and shield connections. It's very interesting: Grounding and Shielding Audio Devices Their bottom line is to connect the shield at both ends to chassis ground, but you probably will have to connect the shields to signal ground which isn't optimal. Either way they recommend grounding both shields.

Wow, learn something new everyday! Thanks.