For the sheer reason of crosstalk attenuation improvement, we give preference to separate grounding.
In a metaphorical sense, the multicore is an extension link for balanced audio lines connecting the audio sources up on stage to the inputs of the mixing desk in the hall. Therefore, it seems quite logical and sensible if we stick to separate shielding on the extended path from the socket connector on stage (via the multicore) to the socket connector on the mixing desk.
Another argument emerges from the ground "potential:" i.e., the point within the system where the screening/s is/are connected to ground, and thus connecting with the PE potential of the location’s lowvoltage power grid. Basically, this should be the input of the mixing console. This is the central point where all the screens of the multicore system are linked to each other. Here we find all screen connections centrally joined, irrespective of the moment, whether on the XLR or the phone plug on the 0 V = ground potential of the internal power supply and thus on PE potential.
With consistently balanced cabling that uses separate shielding for each single line, the XLR input socket on the stagebox is precisely equivalent to the console input to which it is linked. If you connect a condenser microphone here and activate phantom power on this channel, the operating current of the microphone will only flow through this channel.
The separate shielding has even greater benefits when connecting mains-powered audio sources. In most cases we may well assume that down in the hall, where the mixing console has been placed, the PE potential at the power socket will not be exactly identical to the PE potential at the power socket up on stage.
If a mains-powered device is connected with its output to the stagebox, compensating currents may be introduced via the shielding of the audio line, resulting in a classical ground loop. With separate shielding in most cases only this single channel will be affected. Trouble-shooting is relatively easy; inserting a DI box in that line with GND lift (XLR pin 1 isolated) will eliminate the problem. However, if all screens are joined at the stagebox or in the connector and several mains-powered sources are connected, the humming noise can multiply. Not only will potential compensating currents drain off via the console; they will also flow between the connected devices via the common shielding.
One remedy that has spread around among “stage practicioners”, but is hazardous, is when the ground wire is taped or disconnected on devices or power cables. This may lead to temporary sonic success, but is nevertheless EXTREMELY DANGEROUS.
A popular misconception to help save costs is by joining the screens because only the a-wire and the b-wire of a line must be crimped and more audio lines can be used on each connector. However, it is neither the connector, nor the cable, nor the crimping contacts that produce costs there – it is ultimately the technican, who has to braid, solder and connect everything to the screens. Often the screens are then linked to the PE terminal of the connector.
Here, again, the safety aspect is neglected. The terminal on the connector has been marked as CHASSIS = PE terminal by the manufacturer. The connector housing is conductive and must lie on PE potential according to VDE 0100. Doing it any other way is unfortunately forbidden, and we should stick to this protocol. Screens just do not belong at this terminal.
The construction of stagebox and multicore systems is not just a matter of pure signal transmission. After all, you don't want the technology of setting up at a gig have to be so frustratingly complex that it takes all the joy out of public performance for the many amateur musicians and weekend warriors who are simply trying to show off their talents and be more social. Here it’s about tough real-world gigs that just won’t allow you to go trouble-shooting for hours. This could easily be avoided with careful planning and spending a little more money.