To lift pin 1 or not

[Matador]

Seems like there is some confusion from using shield and pin 1 interchangeably.

The shield of a properly made XLR cable is connected to the connector housing on both ends, which will be connected to the jack on chassis, and always have a chassis ground connection on both ends. The XLR pin1 will not be connected to the shield in a properly made xlr cable (although you will find it in some).
This poll was about lifting pin1 from chassis, which will not lift the shield. 
Or maybe I'm missing something?

Is this true?  Google is awash in a plethora of connection diagrams that dispute this:

From Genelec's site:

From Wikipedia:
EIA Standard RS-297-A describes the use of the XLR3 for balanced audio signal level applications: ... Pin 1 = 1 Chassis ground (cable shield)

From Rane's article:

All of these imply that pin 1 is verily the cable shield.

I think the discussion always seems to break down around the right/best way to bridge "chassis" and signal ground.  The Rane article I think addresses this adequately, as I always had the idea that signal ground was always relative to the bit's "inside the box".  Standing outside two pieces of balanced equipment, if box A thinks ground is +40V, and box B thinks ground is -40V, the common mode differential point will be different but in theory the devices transmit audio fine between them.

However the reality is that having a large differential between signal grounds eats into signal headroom (we're always limited by the power supply rails), so having signal ground within reason from box to box is a necessity, and thus we should tie signal ground to chassis ground at one point (and 1 point only) within each box (as Rane shows here):

 

[Speedskater]

Seems like there is some confusion from using shield and pin 1 interchangeably.

The shield of a properly made XLR cable is connected to the connector housing on both ends, which will be connected to the jack on chassis, and always have a chassis ground connection on both ends. The XLR pin1 will not be connected to the shield in a properly made xlr cable (although you will find it in some).
This poll was about lifting pin1 from chassis, which will not lift the shield. 
Or maybe I'm missing something?

In the real world the cable's shield is not connected to the XLR shell (only to pin #1).  The reason being that in the uncontrolled real world, the shell could come in contact with other metal items (these items might have nasty voltages & noise on them).

In an ideal world or a permanent installation, also connecting the shield to the shell would be OK.

 

[dmp]

Is this true?  Google is awash in a plethora of connection diagrams that dispute this:

Maybe I was wrong - I'll have to look at some cables.

 

[JohnRoberts]

I think the discussion always seems to break down around the right/best way to bridge "chassis" and signal ground.  The Rane article I think addresses this adequately, as I always had the idea that signal ground was always relative to the bit's "inside the box".

it has literally been weeks since I posted this but one more time, I really hate the term "audio ground". This only has relevance for non-professional single ended consumer interfaces.

With professional interfaces, outside the box we have audio + and audio -, inside the box we can have multiple single ended circuit blocks each with their own local audio ground, but thinking in terms of a single audio ground is bad for signal integrity since just like power grounds it will not be a single voltage, but will be different depending where you measure it.

Standing outside two pieces of balanced equipment, if box A thinks ground is +40V, and box B thinks ground is -40V, the common mode differential point will be different but in theory the devices transmit audio fine between them.

I once encountered something like 40v AC between the FOH power ground and the stage power ground. This was a less than professional venue on some guys property. A nice natural amphitheater, but lousy mains wiring.

You can configure active solid state inputs that can tolerate large common mode voltages but there is always a trade off. Like padding down the raw input before subtracting the CM voltage, then boosting it back. This will cost S/N and even CMRR. Generally if you see that much voltage in your own venue, you don't tolerate it, you fix it. When the show must go on audio transformers can help git er dun..

However the reality is that having a large differential between signal grounds eats into signal headroom (we're always limited by the power supply rails), so having signal ground within reason from box to box is a necessity, and thus we should tie signal ground to chassis ground at one point (and 1 point only) within each box.

The local signal grounds that I refer to all need to connect to power ground, while audio signal is passed between these different local ground blocks as audio + and audio - using differential amps. 

JR

 

[Matador]

Is this true?  Google is awash in a plethora of connection diagrams that dispute this:

Maybe I was wrong - I'll have to look at some cables.

What I have seen on Neutrik XLR panel connectors is actually a 4th connection:  typically a brass screw lug.  I had always assumed this lug was to have access to the metal shell of the jack in cases where the jack was screwed into a non-conductive panel.

 

[abbey road d enfer]

From Genelec's site:

Genelec are utterly wrong for their unbalanced connection. Since their inputs are balanced, they should take advantage of it and use a balanced connection and tie pins 1 & 3 at the source's end. In fact, that should be the case already in the mixer or monitoring system. If the mixer's output was on TS jacks, a balanced cable should be usd, with shield and cold tied together in teh jack.

 

[Matador]

it has literally been weeks since I posted this but one more time, I really hate the term "audio ground". This only has relevance for non-professional single ended consumer interfaces.

Understood, but we are discussing concepts, not implementations.  Whether or not a piece of equipment has 35 "references that may or may not be called grounds" or a single 0V reference ordained "ground", we still need to make sure that pin 1 hygiene is good.

 

[abbey road d enfer]

Do you connect pin 1 on the output XLR to chassis?

I don't...well, in fact I connect pin 1 to chassis via a parallel RC circuit that prevents high currents to develop in case of differential voltage between grounds, but presents a low impedance to EMI/RFI. I'm in full agreement with Bill Whitlock on that subject.

Argument for is described in the Rane Grounding and Shielding article as "The Absolutely Best Right Way to Do It"

I beg to differ with Rane on that, and with Muncy, Ott... The infamous Pin 1 problem sees the problem only from the angle of rejection of EMI/RFI. There is more to it. A vast majority of sound installations are mobile, and that creates a number of constraints. I don't like the idea of having heavy current developing into the shield in case of high differential ground voltage. This problem has been competely overlloked by them. Whitlock put it right.

 

[Matador]

Abbey, what exact network do you use (e.g. the values)?  Is this something on the order of 100pF in parallel with 100ohms?

I would also assume you would disagree with Rane's idea of a ferrite bead between pin1 and chassis (e.g. this would still allow large DC currents to flow through the shields)?

EDIT: and just to be completely pedantic, you mean via a parallel RC circuit that prevents high currents to develop in case of differential voltage between different chassis, right?

 

[abbey road d enfer]

I did once wireing for audio part of a one big international TV company using their standard project (about 10000 analog and AES wires in several rooms). The shields was connected only on one side, only in rack grounding bars. 

I can't support the idea of lifting the shield at either side. Big antennas.
Some recommend lifting the shield at one end of an XLR cable. This is just ridiculous. Mic cables must ensure ground continuity, particularly if the mic is phantom-powered. I can't imagine having two sets of XLR cables, mic cables with continuity and other cables with one end lifted.
An audio system must be capable of operating at full performance with grounds connected in any combination.

 

[abbey road d enfer]

Abbey, what exact network do you use (e.g. the values)?  Is this something on the order of 100pF in parallel with 100ohms?

Generally 100R and 0.1uF. Low ESR, low inductance and short stout PCB tracks.

I would also assume you would disagree with Rane's idea of a ferrite bead between pin1 and chassis (e.g. this would still allow large DC currents to flow through the shields)?

I haven't seen that but it seems a solution to solve a non-existant problem. The Neutrik EMC protected plug, or more simply a ferrite bead around the cable are more relevant though.

 

[abbey road d enfer]

EDIT: and just to be completely pedantic, you mean via a parallel RC circuit that prevents high currents to develop in case of differential voltage between different chassis, right?

Correct; just as JR mentioned "40v AC between the FOH power ground and the stage power ground"

 

[dmp]

I don't...well, in fact I connect pin 1 to chassis via a parallel RC circuit that prevents high currents to develop in case of differential voltage between grounds, but presents a low impedance to EMI/RFI. I'm in full agreement with Bill Whitlock on that subject.

Generally 100R and 0.1uF. Low ESR, low inductance and short stout PCB tracks.

Excellent contribution. I added this as a category to the poll.

 

[JohnRoberts]

It is clear that audio shields/ground-drain wires are not sized to provide adequate safety ground bonding between chassis, but so what...? I'd rather melt a few audio cables if there's a hot chassis somewhere. If the audio cables draw an arc while plugging in, you might have a larger problem than audio hum..  I appreciate that I am not in the trenches day to day having to deal with large circulating ground currents, so do what you must.

The floating one shield end and cap coupling it is an old hybrid ground scheme practiced in high RF environments, so you can maintain the behavior of a single point ground system for LF, but a highly parallel low impedance ground for RF frequencies. While developed for RF facilities, the increased amount of environmental RF we now experience makes it logical for wider use (if you MUST float a ground). 

Another old trick for lifting grounds while still providing a safety ground path, is to use something like the popular 100 ohm and .1uF disc cap, but also throw a diode bridge between the two grounds. This will look like a high impedance for less than 1V of ground potential difference, but will conduct above that, and if you size the diode bridge large enough, it will hang when hit by a mains power fault long enough to take out the breaker. I talked with UL about approving this to provide primary safety ground bonding inside products, and they were receptive from their preliminary tests, but I couldn't justify spending the tens of $k to open a case file and get it fully vetted by them. which would end up testing and specifying UL approved diode bridges for this application and the like. 

I decided against using the diode trick, because it was cheaper and easier to just hard bond the chassis to mains ground and use proper input differentials for audio signals.

JR

 

[ricothetroll]

if you size the diode bridge large enough, it will hang when hit by a mains power fault long enough to take out the breaker

Why would you need the audio cable shield to play the safety earth role ? Shouldn't the ground fault circuit interrupter take out first in case of a hot chassis ?

 

[JohnRoberts]

if you size the diode bridge large enough, it will hang when hit by a mains power fault long enough to take out the breaker

Why would you need the audio cable shield to play the safety earth role ? Shouldn't the ground fault circuit interrupter take out first in case of a hot chassis ?

Not everybody has RCD/GFCI protected power. Product design for human safety generally relies upon a chassis grounded to safety ground. Not every product has 3 wire line cords and direct path to service ground. Note: in the US ungrounded products require double insulated mains circuitry.

How about a power supply fault in a 2 wire line cord product? That DC path via pin 1 could offer some bonding for the fault current, and while not enough to clamp the voltage as low as a true safety ground, surely enough to trip the RCD/GFCI before melting your wiring.

Like I said, I rather melt an XLR cable than a musician.

JR

 

[abbey road d enfer]

Like I said, I rather melt an XLR cable than a musician.

JR

Depends who.. :

 

[JohnRoberts]

Like I said, I rather melt an XLR cable than a musician.

JR

Depends who.. :

I know you are joking, or think you are, but one of my fears as a product designer is ever being directly responsible for a customers injury or death. In hindsight looking back at my old kit business, I would never let inexperienced customers do their own mains power supply wiring these days. Now we have lumps, wall warts, and self contained PS modules. 

While an employee at a major manufacturer I actually disobeyed an immediate supervisor who told me to submit a design for approval at a second easier test agency after it already failed a stricter test. I refused and had the transformer redesigned to pass the stricter safety test.

Life is precious, even lowly musicians.

JR

 

[abbey road d enfer]

I know you are joking, or think you are, but one of my fears as a product designer is ever being directly responsible for a customers injury or death.

I know the feeling; I live with the sword of Damocles for 35 years now, even more so with most of my stuff being flown over the public. Mechanical safety regulations are even more stringent than electrical safety IMO. I've learnt to live with that and I can sleep at night.

 

[ricothetroll]

Got it, thanx for your answer !