Rob Flinn said:
With that in mind would it be better to fit an RCD on the 115v secondary rather than a Ground leakage type ? That way there is no need for the breaker to be tied in with Earth for it to trip since presumably it doesn't need references to Earth or anything on the primary side to operate.
The GFCI (while perhaps poorly named) does not measure ground leakage or need a ground connection.
As PRR just shared the line and neutral windings (perhaps only one turn each, in opposite directions) create an equal but opposite magnetic flux in a magnetic core that nulls out, as long as Line current and Neutral current are exactly equal. Any imbalance between the two currents shows up as a magnetic flux that can be detected with another very sensitive winding on that same core.
If the line current does NOT exactly equal the neutral current that is "the" problem. At >5mA imbalance it disconnects power.
In fact it does not have to leak to line to ground, it can leak line to neutral somewhere upstream of, or outside the local current sensing loop. You can even trip a GFCI with a bootleg ground to neutral short, where some of the neutral current flows into the ground and creates the imbalance. (bootleg neutral to ground shunts are also code violations). All the GFCI cares about is does the current leaving the protected outlet exactly equal the current coming back in.
I put two GFCI outlets in my house (bathroom and kitchen) replacing ungrounded 2-pin Edisons, with ungrounded 3-pin GFCI receptacles. The GFCI works fine without a ground connection, and this is advised for old 2 circuit wired houses (like mine) for rooms where water and electricity can be present (this morning I put a GFCI in my laundry room.)
By the way as far as I'm aware Neutral in the UK is tied to ground at the Power station. However it can have a bit of induced a.c on it by the time it gets to ones house.
US code bonds ground to neutral one time at the power drop or fuse box.
Power distribution for most US residences comes from a center-tapped transformer winding. The step down transformer steps down the high distribution voltage on the power poles to 230VAC total or two opposite polarity 115V legs (wrt the center-tap). This center tap is also bonded to neutral and ground at the panel/fuse box. 115v branches grab one leg or the other, 230v appliances bridge across the entire winding (230V is two hot legs and no neutral, 115V is one hot and neutral).
It is a code violation to connect ground to neutral anywhere else inside the building, while there are code variations for sub-panels that get treated like a separate power drop.
Don't over think this.... your RCD will probably work if the secondary of your step-down transformer is bonded back to neutral, and the leakage current goes anywhere else other than that same neutral (which is inside the RCD current loop).
Of course adding an extra GFCI in series with the floating step-down secondary winding will also work.
JR
PS: I have seen specialty line cord plugs that sense for current in the ground, but this redundant with GFCI/RCD.
PPS: About 2 years ago I developed a specialty back-line power protection device for musicians to use on stage with dodgy power amps and perhaps dodgy energized microphone grounds (I probably talked about it here). In addition to a GFCI, with touch switch to turn it on that sensed for correct line/neutral polarity, I also detected for current flowing in the safety ground from an external power source. This way I protected the musician from 1: mis-wired stage outlets, 2: legacy (old) guitar amps with bad stinger caps and energized chassis, or 3: energized mic grounds from mis-wired console power outlets.
I abandoned this as too expensive/complicated for a customer base that is more comfortable just cutting off ground pins from their plugs, and dealing with UL approval that would probably take several thousand dollars worth of convincing for a product I would only sell two a year of to cheap musicians.
