I am trying to square this up with a public supply, in which the LV end has either one or two conductors at earth potential, which is ensured by earthing at the transformer.

So is the output of the inverter any different from the output from a public transformer. Granted the cables are short, but they are also short if you live adjacent to a transformer.

If the PCE has an internal impedance of 5 Ω to 6 Ω, then ignoring whether this is Ze or not, how can any Zs be satisfactory?

So is the value of resistance of the electrode not to be considered in a typical earthed generator configuration ?

Yes, it's to be considered - as I recall BS 7430 recommends max 20Ω (although many seem to be on the opinion that significantly higher values can be satisfactory in some circumstances) - what I'm saying that's got little to do with Zs or normal ADS though in a TN system.

The problem with faults to true Earth is the resistance of the fault - i.e. the resistance of the soil that's in contact with the exposed live conductor - we just don't know what it's likely to be - it's likely to be many kΩ - so even if you could calculate something, the numbers are unlikely to work even for RCDs. That's why buried cables with any chance of damage need to have a concentric c.p.c. (e.g. armour).

   - Andy.

I was imagining a mobile generator on standard rubber tyres (like a car). If you touched a live conductor when standing beside the generator, you would not get a shock because there is no circuit.

Yup - protection by separation. Often used with small generators (e.g. ≤ 3kVA) especially when feeding a small system with a limited number of Class I items.

However, the situation becomes different if the generator (along with the chassis, etc.) is earthed. So why do we earth a generator please?

Separation is only good for small systems - larger systems have a couple of disadvantages for a separated system. Firstly long cable runs can mean the system goes get referenced to Earth - via stray capacitance (or even imperfect insulation) to Earth (BS 7671 set a limit of 500m of cable in total, and 100,000 Vm) . Secondly while separated system are only OK on 1st fault - as there's no disconnection on 1st fault, a 2nd fault can then occur which can be fatal - and larger systems will typically have more faults than smaller ones (all else being equal). There's also an EMC issue that some equipment with filters etc, like the PE to be connected to the same system as the live conductors, so they can actually divert the unwanted signals somewhere.

So all in all, larger systems tend to go down the TN route instead.

   - Andy.

Thanks Andy. How about let’s say HO7RNF generator tails on the surface. What happens if they get cut and exposed copper touches the ground?

Surly the path  of the rod back to the star point needs to be low enough to operate the protective device. 

Also, in regards to three phase generators, if one of the lines comes into contact with mass of earth ( True earth) how does the value of the electrode influence potential neutral inversion ? 

Although slightly of topic , this incident is still somewhat relevant. Recently, I became aware of a 630mm single neutral conductor from a transformer to a store being severed and stolen. Remarkably, the electrical supply remained uninterrupted, and the theft was only discovered upon review of CCTV footage. The Distribution Network Operator subsequently visited the site and isolated the supply. I am astounded by the expertise of these thieves, who specifically targeted the neutral conductor.

How about let’s say HO7RNF generator tails on the surface. What happens if they get cut and exposed copper touches the ground?

All depends on the soil resistance and the contact area. As a for instance, let's imagine that the soil is of a type where a 4' x 3/8" rod (so surface area about 30,000 mm²) gives a resistance of 200Ω - and say the gash in the cable exposed say 20x1 mm of conductor (20mm²) - then as a very simple approximation you might imagine that the fault would gave a resistance about 1500x higher than the rod (30,000/20) - so around 30kΩ.. (All very rough, if the soil eas dryer near the surface it could be a lot higher, of if in a puddle it might be lower) Anyhow 230V and 30kΩ would mean a fault current of less than 8mA - so nothing trips, not even a 30mA RCD,

Cue comments about choosing a cable that's suitable for the environment. RN sheathed is pretty robust and isn't easily cut into by the usual blunt trauma - but if sharp implements are a credible risk and other precautions aren't sufficient, then consider armoured (or at least braided) cables instead.

  - Andy.

or to put it another way, ADS works when you can have a fault of negligible impedance - i.e. a fault directly between metallic parts, so you can then know the overall earth fault loop impedance, and calculate on that basis. That's really the essence of "Indirect" shock protection (as it used to be called) - faults to exposed-conductive-parts. Other problems - e.g. due live conductors becoming accessible - is closer to what used to be called direct contact and there's little ADS can do for that. It's either a case of doing your best to make sure that basic insulation remains intact (e.g. by physically protecting the cable) and/or by additional protection (e.g. 30mA RCD) - but even then you don't expect the fault to disconnect immediately it occurs - just when someone is gets a large enough shock for it to be potentially fatal.

  - Andy.

AJJewsbury said:

Separation is only good for small systems

More importantly, systems in which:

1. There is, ideally, no more than one item of Class I equipment; and
2. There are no interconnections between protective bonding or functional bonding circuits in equipment by either direct link such as USB. HDMI, or via capacitive coupling

With regards 2. above, note that wired Ethernet requires a static discharge path to PE (which in some equipment is sent to line or neutral of the supply, which is not strictly in the Ethernet standards) via a suitable resistor/capacitor network.

The above is why IT or separated 'island mode' supplies are strongly recommended against for domestic systems in the IET Code of Practice for Electrical Energy Storage Systems, and are prohibited by MCS standard MIS 3012.

AMK said:

Surly the path  of the rod back to the star point needs to be low enough to operate the protective device. 

Which could be RCDs even in TN-S systems - see IET Code of Practice for Electrical Energy Storage Systems.

AMK said:

Surly the path  of the rod back to the star point needs to be low enough to operate the protective device. 

Which could be RCDs even in TN-S systems - see IET Code of Practice for Electrical Energy Storage Systems.