Just a sight caveat to that AJJ.

I`ve always taken the view (well for yonks) that an RCD is really a must for all normal duty socket outlets on the basis say you have a few dedicated skts RCD protected for outside use and they often get tied up for something else so one of the handier to use more internally sited sockets get used for that outside work. A practice so common I`d say "all sockets are equal in the eyes of RCD" or some such first approximation.

One particularly stood out to confirm my opinion some years ago.

Second floor flat,

All socket circuits RCD protected.

Cooker circuit non RCD protected, the cooker switch unit  had a socket on it. Extension lead plugged in and thru the window where lawnmowing and outside drilling was being done. Oh dear!

There is a great deal of confusion about RCD types and I'm afraid the regs book provides little assistance. However:

There is no regulation of how many protective devices may be present on any circuit. This means that you may have 6 levels of fusing, but as long as one of them matches the circuit parameters (cable, load etc) then it is fine, eg a ring circuit with plug fuses and another inside an appliance or extension cable.. The same applies to other circuit protections such as RCDs, typically a type S and something quick. If there are two RCDs in series, as long as one of them is rated to provide the circuit protection required, then there is no problem, although both may trip but at least one will. This means that a charge point with built in type A or B or F or something new, may be fed from anywhere, whether it has another type which may not provide protection needed by the load, in this case the EV.  Remember you are only providing additional protection for the EV, and it is only needed when the vehicle is charging. At other times a type AC will provide additional protection because there is no DC on the circuit.

This problem has partly come about because electric vehicles use the Earth conductor for signalling with DC, and it looks as though this set off the whole thing with the IEC, and the exact definition of DC is open-ended, as shown in Table 55.3. You should note that there is no figure for the amplitude of these pseudo DC signals, but there seems to be an assumption it is more than 6mA, and secondly that this will prevent tripping of type AC, and increasingly other types. I have done a lot of testing on this and I have a type AC here which is unaffected by 100 mA of pure DC (Real DC), and works at a similar level with the other waveforms. The problem is that some others do not, and so we don't know, and the manufacturing standards are not very clear either, patrticularly when there is a large static AC current and an additional protection type leakage which should cause a trip. Note we always test RCDs completely unloaded.

I hope that lot is helpful, NAPIT is not, its book is in many places inadequate, it should not be used as an installation standard, and any competent inspector does not need it anyway.

Yes you are correct. But there are many installed in a shared rcd. Not ideal and not something I would do but I wouldn't consider it dangerous.

There are many things that should be done that are being dropped in the race to the bottom on pricing, but no one is enforcing the rules so a fully company ev install is unusual fro  what I have seen.

Personally I always try to push for a dedicated rcd but the majority don't.

A circuit which has mechanical protection or in a wall deeper than 50mm do not require RCD protection also sub-mains to out buildings as this is not classed as a final circuit but distribution. 
so in this case it’s SWA does not need RCD protection and the RCD is incorporated inside the EV charger. If there was no RCD inside the EV charger then Type A is required, that’s my understanding. 

The thing with NAPIT Codebreakers is having found the reference to the Wiring Regulations it’s then actually a good idea to open BS7671 and read what it says for yourself, as with any other guidance.

Guidance is what it says it is, a guide.

If you decide that a circuit or particular appliance needs a RCD that can deal with DC leakage and then leave an existing AC RCD that is required for a different purpose upstream then you may have well rendered that upstream RCD inoperable, so effectively you might as well have taken it out and thrown it over a hedge.

Likewise if you need a RCD that needs to deal with DC leakage and fit or leave a Type AC RCD you cannot be certain it will work, even after testing.

I presume the point being made in Codebreakers is if you need a RCD, it has to be suitable and fit for purpose, otherwise it’s useless.

BS7671 is not a design guide, neither is it an instruction book nor an installation manual. 

• It only tells you the expected outcome, in in this case that the installation is protected by RCDs that will function, so although the regs don’t tell you precisely what RCDs to use, not selecting the correct RCD is not acceptable.

OlympusMons said:

531.3.3 "....RCD Type AC shall only be used to serve fixed equipment, where it is known that the load current contains no DC components."

No DC components, not just up to 6mA.

so can't be used on (new or altered) socket circuits.

That probably nails it.

As OlympusMons  said 531.3.3 RCD Type AC shall only be used to serve fixed equipment, where it is known that the load current contains no DC components.

You cannot say that sockets or fixed appliances need a Type A RCD then have a Type AC upfront of it.

Apparently this is the biggest issue found with EVSE installations, which isn’t actually surprising.

I quite see all your comments Andy, but you have missed the subtly covered up error with this idea. I will explain:

First are we providing additional protection or Earth fault protection? In my view it is additional protection because we have the normal Earthing system to cope with Earth faults (I am excluding TT installations for the moment).

The question is then "where is the DC"? The only equipment which could have an actual DC current imposed on the mains waveform is possibly an EV that has this imposed deliberately, but this current is isolated from the mains supply, and for this to matter is probably false due to Kirchhoff's law.

The only way that an RCD can detect anything using the standard differential transformer is that there is a direct imbalance in the phase and neutral current, and the only way to stop this working is to have sufficient alternative magnetic flux that essentially saturates the core, or at least significantly reduces the available permeability so reducing the sensitivity. The list of faults in Fig A53.1 (note I referenced the wrong one above now edited) may introduce an asymmetry in the mains current drawn by each half cycle, but is this DC? The point here is that the net field in the transformer during either half cycle is zero unless the currents in L and N do not balance, and as you know RCDs trip with a single cycle with imbalance although this may need to be repeated more than once (Remember the 0/180 selection on RCD testers?).

The question is now what does Fig A53.1 actually mean? Is the indicated fault path actually an accidental contact with a live conductor by a person, which I doubt, or some kind of fault in the electronics? I doubt it is a person because these terminals are enclosed in all appliances, so it must be a double fault although not shown as such, so that a person becomes the only Earth connection and a leakage fault is also present. In principle I don't care if an appliance leaks to Earth a little, it is not fundamentally dangerous, and if it leaks a lot will pop the CPD.

You may disagree with my discussion above, which is fine, but you need to look at the actual operation in an analytical way and this is difficult, but I would like to hear the argument fully explained as probably would others.

It is alleged in this table and associated regulations that electronics means DC, and yet the only diagram where the Earth current could actually be DC is number 7 because of the smoothing capacitor, probably being several microfarads so that there is not a significant AC component to the leakage. Please consider the live and neutral currents carefully to see if there is a difference.