I am not disagreeing with you Graham, just trying to air the subject as fully as possible.


The next question may be slightly painful to some, and that is: Who in a suitable position to understand the problem signed up to accepting standards which are basically faulty in concept? There appears to be no suitable way in an urban environment where an electric vehicle can be guaranteed to be free of the PME system, even if the supply is "TTed" with a local electrode. So quite simply the now common problem of a lost N connection can make a car significantly dangerous however many RCDs of whatever type are in place. The answer is that we now need to monitor electrode to neutral voltage and disconnect the vehicle completely, including the Earth and control wires, should this voltage exceed some value, say 55V RMS for discussions sake. This disconnection needs to be permanent needing a reset back to the no fault condition and substantially instant. Such a device is fairly simple and cheap to manufacture, but we don't have a requirement to fit one. Instead we have all kinds of very expensive RCDs which do not provide anything like the same level of protection. So lets make a new standard, to incorporate such a device in every charge point, which fixes the problem for good.

Chris Pearson:

gkenyon:

BS 7671 does have another solution in Annex 722 for separation of the vehicle from the installation, if pockets are deep enough.

And still this thread rumbles on.


Graham if you had a PME supply (could be single or three phase) and deep pockets (deep enough to have bought an EV in the first place) which method would you choose please?

Average semi with single-phase supply, I'd rather connect to the PME earth if at all possible, so I'd choose 722.411.4.1 (iii), (iv) or (v), with the following commentary:

• My first choice would be a (iii) device, but for a single-phase supply a measurement earth electrode would be required. This may not always be practicable, as there has to be at least 2 m separation from this electrode buried metalwork connected to the PME system (see Annex I of 4th Ed CoP).

• Following that, I would choose either (iv) or (v) device.

With either option, of course, the first time I used the product I'd want to do some due diligence on the product.


But overall in a small curtilage single-phase property, I think it's better to connect to PME using one of these devices, than try and make a separate TT system, given all the attendant issues of separation and the risk of striking buried services ... and the problems in future maintaining the separation.


If I was going to install V2G system, then that's a different story, as the 722.411 (iv) and (v) devices are not available - but again in this case I'd prefer to use 722.411.4.1 (iii) device with earth electrode than a separate TT system.

Please also clear up what may seem to be an elementary question. I understand the need to separate the earthing arrangements of different systems, which could lead to difficulty placing either a TT earthing rod or a reference electrode. So does the separation have to be from a PME electrode (in my case, I assume that would be a street lamp by the corner of my property) to an installation electrode; or does it have to be from a service cable to an installation electrode?

 

The separation has to be between the separate TT installation earth electrode, and any buried uninsulated metalwork connected to the PME earthing system.

In the spirit of debate still, why does a TT system (or voltage monitoring electrode for a 'unicorn' device) have to be separate from the influence of a PME system? Someone touching the car isn't going to be standing on some theoretically perfect 0V plane, they're going to be stood on the same ground the car is standing on (or at least within a 1m or so of it). So if there are a few tens of volts from a PME system (or from any other source) under the car, surely it's better for the EVSE c.p.c. to be at that same voltage or at least as close as we can make - so we're minimising the voltage difference between the car and the ground it's stood on - rather than some perfect 0V which is only represetantive of the ground potential half a mile away. Isn't that the principle of equipotentiality?


So bang in a rod as close as you can to the (centre of?) the parking space (not hitting buried services notwithstanding) and be done.


   - Andy.

AJJewsbury:

In the spirit of debate still, why does a TT system (or voltage monitoring electrode for a 'unicorn' device) have to be separate from the influence of a PME system? Someone touching the car isn't going to be standing on some theoretically perfect 0V plane, they're going to be stood on the same ground the car is standing on (or at least within a 1m or so of it). So if there are a few tens of volts from a PME system (or from any other source) under the car, surely it's better for the EVSE c.p.c. to be at that same voltage or at least as close as we can make - so we're minimising the voltage difference between the car and the ground it's stood on - rather than some perfect 0V which is only represetantive of the ground potential half a mile away. Isn't that the principle of equipotentiality?


So bang in a rod as close as you can to the (centre of?) the parking space (not hitting buried services notwithstanding) and be done.


  

• If the TT electrode is too close (say within 1 m), it's effectively touching PME, so you've not got separation and therefore no point in the electrode. You also need to allow for ground subsidance.

• You might well then want to argue about the person not standing at the "general mass of earth", but sometimes the voltage drops off quite quickly, even a couple of metres, and there's no guarantee the potential at the feet of a person will be that at the TT earth electrode.

• Conversely, and because of the previous point, in small curtilage properties, you might well be standing over some metalwork (say incoming gas or water pipe) connected to the PME earth, and in certain cases, you simply return the  PME touch voltage ... this of course means that going to the trouble of TT in many dwellings is pointless !

• If it's too close, you're not measuring any voltage difference so the device won't ever operate.

• The minimum separation recommended is 2 m because if you are any closer, with ground subsidance etc., you still need to maintain a reasonable distance.

• At that distance, the device should trip at 40 V, not 70 V

I'm pleased to see in this one thread several hobby-horses of mine that have grown in the last year or two for various reasons.  I'm intending to write them up in detail, and will pass on the links here if that happens (by my getting some time during this period of less work activity).   But here are some notes, welcoming comments.


The "PME thing":  yes, I have long taken an international interest in this, and it seems mainly the UK has this strong and legislated worry about particular applications of TNCS systems. The risks are of course considered elsewhere, but I'm not aware of such a number of prohibited applications (e.g. caravans, petrol stations) or special regulations elsewhere.  Some countries don't use TN* or TNCS much (FR, IT, etc).  Some force the use of TN(usually CS) such as the US and Sweden. Those ones admittedly tend to have 'could be balanced' supplies i.e. split phase or 3-phase. That's far from a cure in all conditions, but it does reduce to some extent the probability of harm. It also avoids the chance of reversed LN polarity going undetected for long! In Sweden many cars have engine heaters that connect the chassis to PE.  They're used with no concern about whether it's TNCS - which it almost always is. In older customer-installations the 'C' part (PEN) may be all the way up to a shared N and PE bar in the fusebox for outgoing final-circuits.  I've seen installations even from the 1990s where a cooker is fed from a TNCS supply coming into an individual flat, and stands next to a sink with no bonding of the water to the PE (there was continuity from sink to cooker, but clearly running through the PEN back to something like a building bond or a pump). But the electric shock deaths per capita are lower than in the UK, and I don't find vehicle-related ones for years back.  That doesn't mean I would touch that cooker and sink together without my gloves on, but I realise I'm probably being oversensitive compared to many other risks with roads, food, etc that I don't know so directly about.  EVs could plausibly be more dangerous than just engine-heaters, being in use during all of the year, including times of likely bare-foot walking (which anyway would raise a question about the safety of 70 V ac).  In countries that 'like' TNCS there aren't exclusions for camping sites, docks, etc etc. In the US I know of a few people now campaigning to get home docks permitted to be, basically, TT: there's very strong resistance from the establishment. They were surprised when I pointed out that their proposal is a method from an IEC standard, used exclusively in some countries, and that the method they're forced to use would be forbidden in the UK. (If you don't know of US systems, consider that in most states there is a  MV [medium-voltage e.g. 13.8 kV] multiply-earthed neutral distributed, to which transformer primaries are connected LN and LV neutrals are required to be bonded, except special exceptions for some farms. Imagine the potentials that can arise on PE during MV faults, or from MV load 3rd harmonics, etc., besides from the LV system. It doesn't take many volts to make a part-submersed person holding a ladder or boat-lift very uncomfortable or uncontrolled.)  Funny world.  Sometimes there are technical reasons for regional differences, which here could include that US RCDs [GFCIs] are electronic and voltage-dependent.  Often, I suspect, it's as much or more a historic matter of what fear has come out on top in the traditional compromise of the pro and con of different earthing systems. 


The attention to TN-C-S is interesting too because it ignores dangers in alternative systems: broken flaky old connections to lead cables in TNS systems together with an earth fault (or leakage), or TT-system RCDs that fail or that can't respond to the current from a particular type of fault. Although load current alone won't cause danger in these systems, the first fault of the connection or RCD could go unnoticed for a long time, until an insulation fault or high leakage does happen. Are the details well enough known to justify special measures on TNCS instead of other systems? (Or special measures at all?)  An old paper (Gosland 1950) does a courageous job of analysing risks (for overhead supply) with what data the author could find. It's probably not very useful now, except as fun, but I really liked the author's response to a comment about the uncertainty of input data: "...refer to the fact that these are based on scanty and divergent data, if not assumed. Nevertheless, any engineer attempting evaluation of the relative merits of methods of earthing must have quantities of this kind at the back of his mind, and there seems to be advantage in stating explicit figures, so that the foundations of opinion may be exhibited and discussed."  I share the UK sentiment of fear of TNCS, and implement this fear in my greenhouse (TT).  But when trying to rationalize the fear, it's not clear that it's justified.


This back-to-BS842 (voltage-operated ELCB) method for EV chargers was fun to see when it arrived.  (Does any other country use this in EV charging?)   I prefer the simplicity of having separate detection of protective conductor current, and tripping all conductors (including PE) if that happens. That seems a good idea for any single socket that feeds outside-the-bonding equipment. I have my own implementation for an outside socket, as one of several "unconventional uses" of a cheap 4-pole type-A RCD.  Tripping is equivalent to unplugging. I later found that the principle is established as 'SPE' RCD (switched protective earth) but is only used in portable RCDs, e.g. the inline ones in an EV charger. Presumably, having this on a fixed socket or charger is disliked because "you mustn't switch the PE" is deeply ingrained.  But I think it would make sense more generally, for single items.  The impedance of vehicles to earth would not cause tripping in normal conditions in a properly operating system, from what I've measured (with my engine heater, even in salty snow).  The protection can operate through wet conditions if the chassis attains a dangerous voltage (I only tried 230V, not 50V), or might require current through a person in order to trip in drier conditions: but as it's only a protection for very unlikely situations it's not unreasonable to see it as supplementary protection that reacts to the shock.  I would use it on any earthing system, not just TNCS.


Yes, type AC RCDs do not seem defensible as a continued product. Last summer I reviewed prices in different countries to indicate that there's not a significant necessary extra cost for A compared to AC.  A better steel is needed, but the international prices make clear that this doesn't have much effect compared to the other factors. As long as most people use AC, then A is "special" and likely to cost a lot more. Having a smaller range of different devices is surely more efficient for production, stock and price. In those places where AC isn't available, A costs much the same as AC does in the UK.  There's practically nothing for which AC is justified, now that cookers and heaters and lights and almost everything else contain diodes as about the first thing at the input. It's obvious that type AC is not used properly in the UK, whatever regulations might wish, since almost every ready-made CU has AC only, and many shops have lots of AC and not a single A.  At least, that was what I found a year and a bit ago when I intended to replace someone's old RCD with two new ones on a visit to the UK.  I was amazed: after not finding suitable ones at several shops including screwfix, I got a fairly expensive quotation from CEF.  It's particularly silly when we have TT systems that depend on the RCD for more than 'just' supplementary protection. (I assume that type-B and AFDDs will also get a lot cheaper with more use, but these at least have some clear further circuitry and cleverness rather than the small construction difference of A vs. AC.)

[In case of confusion regarding the last point in my recent comment, about type-AC RCDs:  I don't mean that every heater, cooker or light will contain diodes, but that many now do: induction hobs, electronically controlled panel heaters, inverter-driven heat pumps, LED/CFL lamps, etc.  So a "cooker circuit", for example, is not clearly a plain resistive load.]

• If the TT electrode is too close (say within 1 m), it's effectively touching PME, so you've not got separation and therefore no point in the electrode. You also need to allow for ground subsidance.

• You might well then want to argue about the person not standing at the "general mass of earth", but sometimes the voltage drops off quite quickly, even a couple of metres, and there's no guarantee the potential at the feet of a person will be that at the TT earth electrode.

• Conversely, and because of the previous point, in small curtilage properties, you might well be standing over some metalwork (say incoming gas or water pipe) connected to the PME earth, and in certain cases, you simply return the  PME touch voltage ... this of course means that going to the trouble of TT in many dwellings is pointless !
  
   

I'd like to comment on  that.

If that was really true, then we would have a safety problem getting into the car, even when the charger is not present, as you are suggesting that significant step voltages are  present due to the house earthing in normal operation, and the car is bridging them.

Where there have been I think 3 deaths of humans from exposed step voltages, and more dogs and some famous horses in the last decade or so, as far as I recall,  these have all be associated with damaged or unterminated underground cables exposing live. As regards earthing faults,  There has been 1 plumber killed by  live buried water main that was acting as CPC until he interrupted it.

In contrast there are a some  hundreds  of  lost neutral events per year in the UK, and this compares with other countries that use aluminum armour on their underground cables.

In practice the surface potential appearing due anything buried at all but the shallowest level is not present as a narrow stripe above the object, rather the surface potential  is a sort of diffused average value with ripples due to sources below, and changes gradually.  Surface finish such as tarmac or concrete are significant, hence the preponderance of free-draining gravel at substations.


It is also worth remembering how RCD blinding works - the core of the current transformer saturates in one direction, so once all the magnetic domains have rotated to align with the external field, and no more can move, it behaves more or less  as an air core, rather than a ferromagnet, and does not 'come unstuck' and start to act as a transformer again, except for those parts of waveform where  the sum of the AC and the DC is back in the linear part of the magnetisation curve, i.e. within a few tens of mA of zero.When the AC term is large, the detector side waveform is largely un affected by the DC  (and the secondary voltage is far from sinusoidal.. ), while when the DC is larger than the AC, no signal is  detected. What happens  in between is slightly unclear as it depends on how well   the detector circuit responds to a 'chopped-up' sinewave where half cycles of one polarity are much peakier than the other.  

I suspect that manyt A type RCDs are just AC designs revisited,  perhaps with a slightly bigger core. Certainly the sensitivity to half wave rectified DC is about half that to a sine wave which would be consistent with this.