[rant] Shielding of audio cables with DIN connectors is weird

[pigrew]

I bought a DIN-7 cable for a project, hoping it was a simple straight-thru cable, but it wasn't! Pin 2 was connected to the cable shield.

It seems that audio-related cables do NOT connect the cable shield to the connector shield. The connector shield was connected to one of eight wires, while the shield was connected to pin 2. It seems that this is standard for MIDI DIN-5 cables.

But, why?! If one wants to leave the shield floating, then fine, leave it open on in the equipment on one end (or add ferrite bead or capacitor or whatever, this reminds me of ground/shield debate for USB). But for the love of God, the connector's shield should be continuous with the cable's shield, and connected to a nice Earth on at least one side.

The best reasoning I can find is that they wanted the shield floating for safety, when using two-prong plugs, so that equipment with a hot chassis wouldn't make the hot terminal easy for the person to touch? Or is there some other reason?

(P.S., is this too technical for the general chat section?)

(Image from The MIDI Association

[Someone]

But, why?! If one wants to leave the shield floating, then fine, leave it open on in the equipment on one end (or add ferrite bead or capacitor or whatever, this reminds me of ground/shield debate for USB). But for the love of God, the connector's shield should be continuous with the cable's shield, and connected to a nice Earth on at least one side.

The best reasoning I can find is that they wanted the shield floating for safety, when using two-prong plugs, so that equipment with a hot chassis wouldn't make the hot terminal easy for the person to touch? Or is there some other reason?

Isolating a chassis mount connector from the chassis is more effort, so they connect the cable shield to a specific pin.

[pigrew]

Isolating a chassis mount connector from the chassis is more effort, so they connect the cable shield to a specific pin.

Good point, but I still can't wrap my head around why one wouldn't want a continuous earth shield. I'm stuck in my RF engineering mindset, where solid ground connections are required in order to provide good shielding (and my feeling that modern equipment has so much high frequency noise that shields are needed for EMC compliance). The project I'm working on has few-MHz single-ended digital signals going through a DIN connector, with a floating circuit common, so I was hoping for something connecting the shield to the connector shield.

MIDI is a bad example for EMI-blame since it's an isolated differential signal. I guess all of this will become outdated as things transition to fiber optics...

[ajb]

This is standard for many cables used in the audio and production world, and it's a safety thing.  Consider that you have a cable plugged into a piece of equipment with a wiring fault that makes the chassis live.  You go to plug the other end of that cable into another piece of equipment (perhaps a long distance away, maybe even on a different mains/generator feed), so you take the metal shell in one hand, and put the other hand on that metal rack.  If the shell is attached to the shield, you're about to have a bad time.  If the shell is isolated from the shield (and the other conductors), then you might be about to buy some new equipment, but that's better than being dead.  Well, maybe that depends on who the equipment belongs to and how much it cost, but the important thing is that you haven't been electrocuted!

Once the connector is mated, typically the shell is connected to the chassis of the equipment, and the sorts of signals on the cables where this is typically done aren't high enough in frequency for this to cause a problem, nor are they sensitive enough to outside interference (often being differential, as in balanced audio or RS485, or current loops like MIDI).

[Cervisia]

The MIDI specification says:

The MIDI input circuit is specified to operate with no connection to ground, in order to avoid ground loops. This is good for audio-frequency signals. However, the cable inductance raises the impedance at high (RF) frequencies making the transmitter ground less effective at longer cable lengths. One solution to this is to connect Pin 2 of the MIDI In jack to ground through a small capacitor. The small capacitor provides a low-impedance path to ground for high (RF) frequencies at the MIDI Receiver side while maintaining immunity to audio-frequency ground loops. A capacitor value of 0.1μF is optionally recommended, which results in an impedance of 0.16Ω at 10 MHz; practically a dead short. At 20 kHz the impedance is 79.6Ω, and at 60 Hz the impedance is 26.5kΩ.

(Pin 2 is always grounded at the transmitter.)