I still don't understand how I can tell if the PEN is likely to break when the voltage is above 253V or below 207V. Or will the voltage be higher than 253V or lower than 207V when the PEN is open?

Very generally, the problem is to be able to work out what voltage the PEN conductor (and hence the exposed-conductive-parts of the vehicle) are at relative to true Earth, without having an actual true Earth reference to measure it against. The idea is to use the supply L conductor as a reference (as it's generally 230V/240V above true earth as it's earthed at the supply and and that reference isn't lost in a simple broken PEN event).

A entirely single phase supply is probably the simplest example. When the supply PEN conductor is broken, N/PE conductors on the load side of the break (including the installation MET etc) are pulled towards line voltage by any connected loads and may be dragged towards earth by any connections to true Earth (e.g. bonding to extraneous-conductive-parts). Any voltage on the installation's PEN conductor then shows up directly as a reduction in apparent voltage between PEN and L. So say the MET was at 50V and the supply 240V - L/N and L/PE voltages in the installation would be in the region of 190V.  If that were the only case, the devices could be set to trip at anything below 170V and that would be that.

Many installations, including single phase ones, are fed from polyphase (usually 3-phase) LV distribution systems though. That makes things rather more complicated. The connected loads are pulling the severed end of the PEN not just towards L but in almost a tug-of-war between all three L, 120° apart. If all loads were perfectly balanced the effect of the 3 Ls would precisely balance and the severed PEN would remain at 0V (comparable with omitting the N on 3-phase circuits with no single phase loads). In practice though, loads at installation level are almost never perfectly balanced, so (thinking of a phasor diagram) the severed PEN could be pulled anywhere within a 400V triangle - so any single 1-phase load (e.g. a charge point) could see anything between 0V and 400V between L and N or PE. Hence the idea of taking what's "normal" and treating anything outside that as being indicative of a possible broken PEN.

In the UK, the supply *should* be between 216.2V and 253V (it's a statutory requirement), so normal adding or removing loads and allowance for voltage drop in the supply system should still keep things within that range. There may be some short duration excursions outside - e.g. for starting currents - but they should be limited (and the DNOs have powers to have "disturbing" equipment disconnected from the supply). Likewise routine "faults" (e.g. L-N or L-PE faults within installation) can cause the supply voltage to collapse (or increase if the fault is on a different phase) - but again the duration should be limited (faults in any modern installation are normally disconnected well within 5s). So the "normal" voltage range, plus a time delay (e.g. 5s), should avoid the vast majority of problems.

   - Andy.

I still don't understand how I can tell if the PEN is likely to break when the voltage is above 253V or below 207V. Or will the voltage be higher than 253V or lower than 207V when the PEN is open?

Very generally, the problem is to be able to work out what voltage the PEN conductor (and hence the exposed-conductive-parts of the vehicle) are at relative to true Earth, without having an actual true Earth reference to measure it against. The idea is to use the supply L conductor as a reference (as it's generally 230V/240V above true earth as it's earthed at the supply and and that reference isn't lost in a simple broken PEN event).

A entirely single phase supply is probably the simplest example. When the supply PEN conductor is broken, N/PE conductors on the load side of the break (including the installation MET etc) are pulled towards line voltage by any connected loads and may be dragged towards earth by any connections to true Earth (e.g. bonding to extraneous-conductive-parts). Any voltage on the installation's PEN conductor then shows up directly as a reduction in apparent voltage between PEN and L. So say the MET was at 50V and the supply 240V - L/N and L/PE voltages in the installation would be in the region of 190V.  If that were the only case, the devices could be set to trip at anything below 170V and that would be that.

Many installations, including single phase ones, are fed from polyphase (usually 3-phase) LV distribution systems though. That makes things rather more complicated. The connected loads are pulling the severed end of the PEN not just towards L but in almost a tug-of-war between all three L, 120° apart. If all loads were perfectly balanced the effect of the 3 Ls would precisely balance and the severed PEN would remain at 0V (comparable with omitting the N on 3-phase circuits with no single phase loads). In practice though, loads at installation level are almost never perfectly balanced, so (thinking of a phasor diagram) the severed PEN could be pulled anywhere within a 400V triangle - so any single 1-phase load (e.g. a charge point) could see anything between 0V and 400V between L and N or PE. Hence the idea of taking what's "normal" and treating anything outside that as being indicative of a possible broken PEN.

In the UK, the supply *should* be between 216.2V and 253V (it's a statutory requirement), so normal adding or removing loads and allowance for voltage drop in the supply system should still keep things within that range. There may be some short duration excursions outside - e.g. for starting currents - but they should be limited (and the DNOs have powers to have "disturbing" equipment disconnected from the supply). Likewise routine "faults" (e.g. L-N or L-PE faults within installation) can cause the supply voltage to collapse (or increase if the fault is on a different phase) - but again the duration should be limited (faults in any modern installation are normally disconnected well within 5s). So the "normal" voltage range, plus a time delay (e.g. 5s), should avoid the vast majority of problems.

   - Andy.