XLR Three vs Two Conductor / Ground vs No Ground / To Sheath or Not To Sheath

[Zeotrope]

Can we have a definitive guide from the experts here on what the deal is with the various ways to wire XLR interconnects?

Opinions diverge as widely as possible. I have well respected audio designers telling me that floating ground, 2 cable XLR (ie nothing connected to Pin 1) is best to minimize noise, while others say the opposite.

To ensure we are using the same Pin nomenclature:

 

[Zeotrope]

I have noticed that high-end cables from Stiltech and Luna for example have three conductors.
I just ordered a Mogami 2534 XLR for testing which also has all three conductors wired. This cable is used by professionals especially in environments where there is a lot of RFI/EMI.

 

[Zeotrope]

The opposite approach:
“2 identical cables to PIN 2 [+] and PIN 3 [-] Only on the "send" side [FEMALE XLR] is the negative [-] connected to the STEEL outer shield. So it floats on the "receive" side [MALE XLR].
This works best for both single-ended and balanced connections to avoid creating an earth loop.”

 

[Alrainbow]

To be technically correct there is 4 cond
A shield
A common is used for both positive and negative signal conductors
2 signal conds
There is no industry standards
The shield in my view should only be connected on 1 side.
Source side I think.
Like data wire when it's shielded
What we need to know is what devices to be used
When having cabels made

 

[Alrainbow]

xlr pin diagram - Google Search

 

[Alrainbow]

We need an industry expert pronto
Lol

 

[Zeotrope]

From Tony @ Humble Homemade (a great guy with a ton of knowledge, hope he doesn't mind the quote):

"If a cable is more or less susceptible for hum or buzz depends on a few things, mainly the amount of shielding but also the cross-section of the conductors.

Noise can occur between components as the ground or shield wire will always look for the shortest way to ground. The path with the lowest resistance. Sometimes this is not directly to ground via the earthed mains cable of the device but via an interconnect and then to ground via the earthed mains cable of a different device. The more complex a system (the more devices connected together) the more difficult it is to determine where noise is being generated and which path it follows to ground.

The 3-pins of an XLR cable must correspond with how the manufacuturer of the amps and source equipment have wired them. And if the cable only has 2 pins connected, not three, then it is not a balanced cable. The idea behind XLR is that the shield is spererate from the two signal wires. So basically the 'dirt' stays in the shield and the + and - signal wires are 'clean'."

 

[DaveC]

Industry standard for an XLR cable is a shielded twisted pair or star quad.

The conductors for pins 2 and 3 must be symmetrical, this is very important.

Shielding is a double-edged sword that can, well, shield, but it also adds capacitance and in most cases impinges on the EMF surrounding the cable, which has an audible effect on the signal. Shielding in a home environment may not be necessary, but you still need a heavy gauge connection for pin1. I use either a tinned braided shield or heavy gauge wiring in the same place a shield would be.

Noise occurs between components mostly as a result of the component chassis being at different potentials due to different potentials at the IEC inlet ground, or as a result of leakage currents from trafos, capacitance to ground, etc.. An XLR cable should have it's pin1 connection attached to chassis right next to the in/out jack, and thus current will flow due to ground potential differences, but should affect pins 2 and 3 equally and thus be canceled out. In RCA cables it'll take the ground/return and in some cases it can also use the ground wires in your power cables, this depends on if the IEC ground is isolated. You want heavy ground wires in ICs and PCs, and you see studios using buss bars and strapping chassis to these bars as the lower the resistance the less noise voltage generated as a result of ohms law, V=IR, so reducing R also reduces V.

 

[Zeotrope]

Thanks Dave, appreciate the insights here.

"Shielding in a home environment may not be necessary, but you still need a heavy gauge connection for pin1". So this means an actual conductor connecting Pin 1 at both ends, correct?

In my case, with the DarTZeel 18NS preamp, the connection to the DartZeel 108 stereo amp was via (unbalanced) 50ohm BNC cable (10m). There was no ground noise -- but it's possible that the ground was lifted (DarTZeel's jumper labels are not that clear, as each side (pre and power) has different terminology and I'm not 100% sure the jumpers actually work).

With the Nagra REF Pre and REF mono blocks, the same BNC cable (now terminated with RCA) produced loud buzzing. According to Nagra, the RCA unbalanced is grounded.
I switched to XLR and the buzzing was mostly gone, but is still faintly heard. Now, the Metaxas XLR I am using has no third conductor on Pin 1. And the shield is connected to Negative only on the sending/preamp side. This does not seem like a truly balanced cable, and probably why I still have some noise?

All components are on the same AC circuit and all components are connected to the Shunyata ALTAIRA grounding system. I even tried connecting the Pre and Amps to the same outlet and the noise was the same - it only (largely) disappeared when using the XLR (but it's still an unbalanced XLR).

 

[DaveC]

Thanks Dave, appreciate the insights here.

"Shielding in a home environment may not be necessary, but you still need a heavy gauge connection for pin1". So this means an actual conductor connecting Pin 1 at both ends, correct?

In my case, with the DarTZeel 18NS preamp, the connection to the DartZeel 108 stereo amp was via (unbalanced) 50ohm BNC cable (10m). There was no ground noise -- but it's possible that the ground was lifted (DarTZeel's jumper labels are not that clear, as each side (pre and power) has different terminology and I'm not 100% sure the jumpers actually work).

With the Nagra REF Pre and REF mono blocks, the same BNC cable (now terminated with RCA) produced loud buzzing.
I switched to XLR and the buzzing was mostly gone, but is still faintly heard. Now, the Metaxas XLR I am using has no third conductor on Pin 1. And the shield is connected to Negative only on the sending/preamp side. This does not seem like a truly balanced cable, and probably why I still have some noise?

Pin1 is basically an extension of the chassis that covers/shields the signal wires, I've heard it said that it's not strictly necessary but component chassis really do need to be grounded together. If you've ever hot-swapped RCA cables that make the signal connection before the ground, you've heard the result of not having grounds connected. XLR is different as the signal wires should be equal and opposite in potential but imo pin1 is still required.

Where there are no real standards is IEC ground to chassis connection. This is required for safety unless the chassis is double insulated, so there are restrictions on isolation due to safety concerns. If there is a short to chassis ground you want as little resistance in the ground connection as possible so the breaker trips as intended. But there's probably some amount of isolation, maybe a ground loop inductor, and a few other schemes to isolate IEC ground from chassis. So if the IEC is isolated the pin1 connection will provide a direct connection between chassis, get rid of that and you may experience that buzzing. So, I would guess that a XLR cable that meets industry standard might solve your problem as it'll provide a ground connection via pin1 using a shield, which is generally heavy gauge if it uses braided copper. I would bet having this connection intact will eliminate your buzzing.

 

[Zeotrope]

Thank you, much appreciated!