> An ideal transformer will give you perfect galvanic isolation. A GFI therefore becomes useless on transformer secondaries, because there is no way to complete a circuit through earth back to the transformer.
Yes and no. You declare one of the secondary wires to be neutral and the other to be the “line”, and you connect “ground” to “neutral” in exactly one place. And the GFI monitors the sum of the line and neutral current, but not the ground. And you connect your equipment chassis to ground via the wire that, in the US, is green and is called, in the NEC, the “equipment grounding conductor”. And the GFCI will protect you if you are touching a grounded object with one part of your body and accidentally contact something that’s is at “line” voltage.
If you are using a transformer to provide power to part of a building subject to the NEC, you may be required to do approximately this. And your bathroom receptacles are hopefully GFCI protected and are also fed from a transformer near your house…
> electrostatic shield
You can get a transformer with approximately this, in small sizes, sold for “medical” use, off the shelf at reasonable prices. I have one that I bought from Digikey because it had the right specs, even though I don’t need the shield.
But I haven’t seen one that has fittings intended to integrate it into a Faraday cage. Maybe someone makes it. Admittedly, I have not worked with that many Faraday cages in my life, and I didn’t get to examine the detailed construction of the giant anechoic Faraday cage in which we tested the smaller Faraday cage I was using.
There's no reason you have to connect the secondary to earth at all. This is how a lot of medical equiment (under IEC 60601) and test equipment (under IEC 61010) is designed. If earthing it makes everything worse, just don't do it.
The NEC might or might not have something silly to say about that, but when you have a giant custom-built Faraday cage, I think most people will be understanding that you might choose to call it "equipment" (and thus, for example, claim that 61010 applies) rather than "building wiring" (so not NFPA 70).
You can also earth things in some really weird ways. My favorite was to connect the secondary to earth through a 12V bidirectional TVS diode. Apparently this stops a lot of nuisance issues while still allowing serious currents to flow. Unfortunately, you want a bigass TVS to do this, and those are rather capacitive. So you still end up with RF issues. (No, you cannot use this to make an arc fault detector. Or at least, I couldn't do it.)
Electrostatically shielded transformers are not really stock items. You will usually have to get them custom-built and they will usually end up as expensive, higher-performing C-type cores. Shoutout to Triad Magnetics for being very nice to work with as a custom magnetics vendor on that horrid project even though the project was cursed and we ended up going with someone else (which we should have known a lot earlier and not strung you guys along, but engineering and management had different aims there...), right up until it got cancelled.
Yes and no. You declare one of the secondary wires to be neutral and the other to be the “line”, and you connect “ground” to “neutral” in exactly one place. And the GFI monitors the sum of the line and neutral current, but not the ground. And you connect your equipment chassis to ground via the wire that, in the US, is green and is called, in the NEC, the “equipment grounding conductor”. And the GFCI will protect you if you are touching a grounded object with one part of your body and accidentally contact something that’s is at “line” voltage.
If you are using a transformer to provide power to part of a building subject to the NEC, you may be required to do approximately this. And your bathroom receptacles are hopefully GFCI protected and are also fed from a transformer near your house…
> electrostatic shield
You can get a transformer with approximately this, in small sizes, sold for “medical” use, off the shelf at reasonable prices. I have one that I bought from Digikey because it had the right specs, even though I don’t need the shield.
But I haven’t seen one that has fittings intended to integrate it into a Faraday cage. Maybe someone makes it. Admittedly, I have not worked with that many Faraday cages in my life, and I didn’t get to examine the detailed construction of the giant anechoic Faraday cage in which we tested the smaller Faraday cage I was using.