Remember a XLR balanced interconnect system has no ground connection, it only has a shield connection to the chassis. The fact that the Safety Ground/Protective Earth is also connected to the chassis has nothing to do with the balanced interconnect system.
JMS: I only said the word "ground" because it's typically called a ground lift by many manufacturers. It should be called a pin-1 "bond" but then nobody would know what I was talking about.
A correctly designed unit (one that does not have a pin 1 problem) should not experience any problems regardless of the number of balanced interconnects.
JMS: True, but what do we do about the many thousands (millions?) of active speakers and amps with the pin-1 problem. We can't just throw then all away.
Tony Waldron, a UK large system expert (CADAC Electronics) recommends that all shields on all cables be connected, even in multi-building systems.
JMS: I don't agree with that. Engineers in the UK don't have to deal with how bad our American grounding systems are. Much (maybe most) of our EGC (safety grounds) are contaminated with multi-point bonding to building steel. From what I've read, the UK doesn't have that problem. Hence their engineers don't think about it much.
Maybe people should build a John Windt "Hummer" tester.
JMS: Actually, the Mike Sokol version is much better. It can generate up to 3 volts and 40 amperes of ground loop differential voltage and current.
A double insulated enclosure should not be modified. I would think that low powered double-insulated equipment will become more and more popular.
JMS: True that. But I've run across one of these (a Roland cube guitar amp I think) that developed a line to chassis short, possibly from a nearby lighting strike. That's what I fear most about double-insulated appliances. However, a GFCI on that outlet should protect the meat puppets from electrocution.
BTW: Here's a pic of the business end of the Glo-Melt transformer I use in my Pin-1 hum injector. There's a huge amount of amperage available in many grounding systems, and this will easily produce 30 amperes of ground loop current. As I've noted before, my experiments show a typical ground loop current of around 1 ampere per volt of ground loop differential. I've measured between 2 and 5 volts difference between EGC grounds in American wiring, so that implies up to several amperes of current that an XLR shield has to pass if its bonded on both ends and there's a significant voltage between EGC connections.