RCA interconnect shield connection: to source or receiving end?

[uwotm8]

Warning: audiofool content (18+)

I guess that have been asked like thousand times but I just need to get it clear? so...
We have "balanced" RCA cable structure: two isolated signal wires and separate/external foil shiled tube+drain wire touching it.
Proper structure mean that shield/drain is connected to the ground only on one end, while both signal wires connected to both ends ofc.
But what side is correct for drain connection: source or receiver?
I was sure that drain should be connected to source as it made on my Profigold PGA3000s (cheap but lovely cables) and tried to reverse it once I've read that drain should be on amp end - believe or not, that sounded shomehow worse (audiofool detected for sure). So I left it like marks say - shield-to-source.
Then suddently I got my eyes on AudioQuest cable termination manuals for bulk cables and they straight and forward say to connect shield to receiving end (amp).

https://www.audioquest.com/resource/1132/Chicago-RCA_Termination-Manual.pdf

https://www.audioquest.com/resource/1130/RedRiver-Mackenzie_Termination-Manual.pdf

Just got 2m Chicago factory terminated pair
Thinking about some garage-DIY with Mackenzie as far as bulk version is fairly priced

So... Audiofoolness away, what end for shield connection does science suggest? And why?

 

[MakeMineVinyl]

RCA is an unbalanced interconnect. Both inner conductor and the shield have to be connected at both ends. XLR balanced connections can operate with the shield connected at one end only, and which end can be different depending on who you ask.

 

[ta240]

Theirs has 2 wires internally with a shield around them, so the shield doesn't have to carry the signal. That is funny with Audioquests take on it, everything I've seen says to connect the shield on those to the source side. In reality does any noise picked up by the shield really go to one component instead of both?

 

[uwotm8]

RCA is an unbalanced interconnect. Both inner conductor and the shield have to be connected at both ends. XLR balanced connections can operate with the shield connected at one end only, and which end can be different depending on who you ask.

I decribed what exact RCA structure I'm talking abiout, not sure what is right specific English term for it - "open shield"? Anyway, it's also showed in AQ's PDFs.
As I said I like PGA3000s what I own for actually making sound cleaner, but the question is not about me being audiofooled
If we assume that such open loop works (if you look you'll find that many of such double or triple shielded cables are made within "open shield" concept) where'd you connect shiled to ground?

 

[MakeMineVinyl]

I decribed what exact RCA structure I'm talking abiout, not sure what is right specific English term for it - "open shield"? Anyway, it's also showed in AQ's PDFs.
As I said I like PGA3000s what I own for actually making sound cleaner, but the question is not about me being audiofooled
If we assume that such open loop works (if you look you'll find that many of such double or triple shielded cables are made within "open shield" concept) where'd you connect shiled to ground?

An unbalanced connection using two wires and a shield accomplishes nothing beyond a simple coaxial cable. If you want something approaching cancellation of hum/emi/rfi, then go balanced using XLR.

 

[uwotm8]

An unbalanced connection using two wires and a shield accomplishes nothing beyond a simple coaxial cable. If you want something approaching cancellation of hum/emi/rfi, then go balanced using XLR.

So the answer is "such shield doesn't work", OK, noted

 

[MakeMineVinyl]

So the answer is "such shield doesn't work", OK, noted

It has been argued that the shield acts as an electrostatic shield but the real solution is to use a true balanced connection where the shield doesn't share the signal ground.

 

[uwotm8]

It has been argued that the shield acts as an electrostatic shield but the real solution is to use a true balanced connection where the shield doesn't share the signal ground.

For some reason the one cable that makes some notable difference for me is such fake-"balanced" RCA.
Will see hear how AQ Chicago works, it's interesting.
All my devices are single-ended so no XLR and real balanced connections

 

[KSTR]

So... Audiofoolness away, what end for shield connection does science suggest? And why?

Science suggests that any unbalanced connection should have the lowest possible resistance/impedance for the "return" connection, this lowers susceptibility for "ground loop" type of noise ingress (voltage drop along the shield creating an direct error voltage at the receiving end even if the input side is shorted).

And to reduce magnetic and electric fields coupling into the signal, you want a coaxial cable (geometrical symmetry is key, so a true tubular braided shield with high coverage -- two layers, preferably --, not foil + drain wire) with the shield connected on both ends, to extend the Faraday cage from device A to device B. Shield on one end only does not help for magnetic fields and if the two inner wires are not tightly twisted they form enough loop area to pick up magnetic field.

Twin leads with an outer shield connected to one end does not cut it in both ways.

 

[DVDdoug]

Fortunately, noise pick-up in line level cables is rarely a problem. If you have (audible) hum it's probably a ground loop and the best solution to that is a true balanced connection.

The low output impedance helps to minimize noise pick-up, and the relatively high signal level helps with the signal-to-noise ratio.

Phono pickups are higher impedance and lower level so any noise gets amplified along with the signal and that makes it worse in two ways. But the phono pickup has coil which is "ideal" for picking-up electromagnetic hum (although it too, is shielded) and the cable itself is less efficient at picking-up noise. Guitars are also high impedance and they seem to be even "better" at picking up hum & buzz.

Pro microphones are low-impedance balanced, so although the signals are weak hum is usually less of a problem than hiss from the preamp.

 

[Speedskater]

The "KSTR" post number 9 wins.

 

[uwotm8]

Twin leads with an outer shield connected to one end does not cut it in both ways

But that design is widely and commonly used for decades and by variety of brands from snake oil sellers to cheap "just connect ur s**t" B-category brands

 

[MakeMineVinyl]

But that design is widely and commonly used for decades and by variety of brands from snake oil sellers to cheap "just connect ur s**t" B-category brands

That doesn't make it right. The laws of physics aren't subject to a popularity contest.

 

[wwenze]

The problem with single-ended systems and shielding is that, which secondary ground should receive the EMI depends on individual setup. This requires either measuring and calculating the electrical parameters, or measuring the end result. A concept that goes against audiophilia.

 

[restorer-john]

The other big issues with the RCA single shield at one end and the twin conductors (one for hot, one for gnd) are the following:

The twin internal wires have significant capacitance between them, as they are intended to be as close together as possible for balanced use. Measure it- you will be shocked (not literally).
The resistance of the shield is always much lower than the resistance of an internal core, so the outer/gnds on the connected gear can have a differential (=hum)
Even if you parallel the internal pair and connect the shield at both ends, it's still inferior to a single core coaxial shielded cable for RCAs.

Bottom line: Use coax or single core shielded for RCAs and twin core shielded (or even fancy quad core if you like spending money) for balanced. Forget the dumb ass 'psuedo balanced at one end or not at all' wiring.

 

[mansr]

I'm just going to drop this graph here:

 

[sam_adams]

To follow on with what @restorer-john was saying, the coaxial cables examples below are excellent choices for interconnects. Look at the specs and choose the one with the lowest shield resistance (Some have shield resistance as low as 1 Ω.). Those with stranded center conductors and braided shields are the most flexible. Braid and foil shields should be chosen for interconnects that would be left in place and not repeatedly connected and disconnected as this risks flexing the foil and cracking it which reduces its effectiveness.

Keep in mind that the characteristic impedance of the cable—50 Ω or 75 Ω—means little to us for our frequencies of interest. It is only important at RF frequencies. We are most concerned with the controlled impedance that the precise geometry of these cables provide in the audio band.

Stranded center conductors with braided shields:

Belden-1505F-RG59U 22 AWG Stranded Center Conductor - Double Braid Shield - 6.15 mm

Belden-8219-RG58AU 20 AWG Stranded Center Conductor - Single Braid Shield - 4.9 mm

Belden-8241F-RG59U 22 AWG Stranded Center Conductor - Single Braid Shield - 6.12 mm - Use Canare RCAP-C4F crimp on connector

Belden-8259-RG58AU 20 AWG Stranded Center Conductor - Single Braid Shield - 4.9 mm

Belden-8262-RG58 20 AWG Stranded Center Conductor - Single Braid Shield - 4.95 mm

Belden-8281F-RG59 21 AWG Stranded Center Conductor - Double Braid Shield - 7.72 mm

Belden-9203-RG58U 20 AWG Stranded Center Conductor - Single Braid Shield - 4.95 mm

Canare LV-61S RG59B/U Stranded Center Conductor - Single Braid Shield - 6 mm

Canare LV-77S RG59 Stranded Center Conductor - Double Braid Shield - 7.7 mm

Mogami W2964 24 AWG Stranded Center Conductor - Double Served Shield - 4.8 mm

Mogami W3200 20 AWG Stranded Center Conductor - Double Served Shield - 4.8 mm - Hard to find in US market. EU and Asia may have better luck.

Solid center conductor braided and braid/foil shields (some exotic and expensive to source):

Belden-1855A-6G SDI-RG59U 23 AWG Solid Center Conductor - Braid/Foil Shield - 4 mm

Belden-4505R-12G SDI-RG59U 20 AWG Solid AgCu Center Conductor - Braid/Foil Shield - 5.92 mm

Belden-8281-RG59 20 AWG Solid Center Conductor - Double 95% Braid Shield - 7.75 mm

Belden-8281B-RG59 20 AWG Solid Center Conductor - Double 98% Braid Shield - 7.75 mm

Belden-9273-RG223U 19 AWG Solid AgCu Center Conductor - Double AgCu Braid Shield - 5.38 mm

Stranded with braid and foil shield:

Belden-9907-RG58 Thinnet 20 AWG Stranded Center Conductor - Braid/Foil Shield - 4.7 mm

Belden-82907-RG58A/U Thinnet 20 AWG Stranded Center Conductor - Braid/Foil Shield - 4 mm

Belden-89907-RG58A/U Thinnet 20 AWG Stranded Center Conductor - BraidFoil Shield - 4.2 mm

When attaching connectors, keep the center dielectric as close as possible to the center terminal—1/16 inch (2 mm). Keep the shield as close as possible to the end of the stripped center dielectric, also. This requires skill and patience with the wire stripper and soldering iron.

If you use a connector like this:

Place a spot of solder where it says "Solder here" to ensure the lowest possible resistance to the shield. That shield connector is only making contact through the rolled-over portion of the connector that holds the center insulator in place and can have a relatively high resistance compared to the shield conductor's resistance.

If using a connector similar to this:

Solder the shield conductor as close as possible to the end of the outer shell that holds the insulator and center pin.

Whenever you're soldering any of these connectors, make sure that you heatsink the metal in contact with the plastic insulator for the center pin. Teflon and Poly-Acetal will melt faster than the solder and the pin can get off center and ruin the job very quickly. Use a female RCA connector and one of those third-hand gizmos with the alligator clips on them to keep things in place.

Once done, clean off the flux with alcohol and check with a hand lens for solder balls and splash. You should also check your work for shorts and use a milli-ohm meter to check the quality of your soldering. The final resistance of the cable assembly should be very close to the total DC resistance of the cable length to within a few milli-ohms. If not, try again.

Finally, if you don't want to put yourself through any of this, you can always buy pre-made cables.

 

[Human Bass]

The best rca wires I know is the gotham gac-1 ultra pro. Coaxial with 4 layers of shielding, flexible and no "directional" BS. WBC makes rca cables with them.

 

[uwotm8]

gotham gac-1 ultra pro

As far as it's made of convetional OFC copper (non LGC or PSC+ as AQ cables) I don't believe it can sound good
---
Jokes off, that one seems to be overkill

Solder the shield conductor

I'm OK with soldering basics but thanks

That doesn't make it right. The laws of physics aren't subject to a popularity contest

I'm not that arrogant to be satisfied with saying "hah they're fools with no physics knowledge".
There's too much of one-end-connected RCA interconnects out there and such explanation is not enough - for me at least.
Maybe it's just a ground loop prevention etc etc but there has to be a reason

As I expected that was discussed earlier:

XLR input to amp module connections?

The highlighted portion in yellow is nonsense, right?

www.audiosciencereview.com

 

[uwotm8]

Belden-9273-RG223U 19 AWG Solid AgCu Center Conductor - Double AgCu Braid Shield - 5.38 mm

Silver-Plated Copper Covered Steel
...for real?
List is full of TV coaxials that I'd never use at least because it's simply uncomfortable to work with - unlike typical Canare balanced/mic audio intecronnect