UK Wiring Regulations and Charge Point Safety

[Glan gluaisne]

Moderator comment - the following threads were moved from Rolec EV Wallpod

Thanks @Glan gluaisne to bring this up

I had this issue addressed in my post #21 but was shot down...

Unfortunately, there is a myth that's still being promulgated that the requirements for wiring an EV charge point somehow changed relatively recently. The reality is that they didn't, at least not materially. The requirement for having DC tolerant RCD protection and some form of open PEN protection (for PME installations) dates back to January 2012, when the IET published requirements that were broadly similar to what eventually got incorporated into BS7671. Although those IET requirements didn't have the authority of being a part of the wiring regulations, they were, in effect, given the same authority, as the government mandated that they be complied with for any charge point installed with the aid of grant funding.

The reality is that the majority of charge points that were installed in the early days, from when grants became available, until BS7671 was amended to include Section 722, were not safe and compliant with the requirements. Some manufacturers chose to just ignore the need for DC tolerant RCD protection, and unfortunately it seems that few electricians actually understood that this was a requirement, so didn't question it. The result is that we have a lot of older charge points that have never met the safety requirements from the day they were installed, and some, including a few electricians who should know better, aren't really picking up on this. It's something that would definitely be a C2 on a periodic inspection, though, so we should be seeing some of these non-compliant installations being picked up through the ten year EICR process.

Edited: corrected typo

 

[arg]

Unfortunately, there is a myth that's still being promulgated that the requirements for wiring an EV charge point somehow changed relatively recently. The reality is that they didn't, at least not materially. The requirement for having DC tolerant RCD protection and some form of open PEN protection (for PME installations) dates back to January 2012, when the IET published requirements that were broadly similar to what eventually got incorporated into BS7671.

This is not accurate. There were material changes to the RCD requirement in BS7671:2018 (published Jul 2018, effective Jan 2019) and to the open PEN precautions in BS7671:2018+Amd1 (published Feb 2020, effective Aug 2020).

The history of specifications specific to EV chargepoints is:

• Code of Practice (1st edition), published 2012. This provided guidance to supplement the wiring regulations currently in force (17th edn, BS7671:2008) which had no specific regulation for chargepoints, though of course contains lots of general principles which are applicable.
• BS7671:2008 Amendment 2, published August 2013. This was a stand-alone amendment which inserted a new section (722) to contain regulations specific to EV chargepoints. This broadly codified the same approach as had been recommended in the CoP.
• Code of Practice (2nd edition), published 2015. Contains guidance in interpretation of the newly-introduced regulations, particularly relevant where the regulation talks about "reasonably practicable".
• BS7671:2008 Amendment 3, published Jan 2015. This was a complete re-print of the book to incorporate minor amendments throughout, however section 722 remained identical to the version that had been issued in Amendment 2. It did however bring the Amd2 changes to a wider audience - all electricians would have to get a copy of the Amd3 "yellow book" whether they dealt with EVs or not.
• BS7671:2018 "18th edition", published July 2018 as a complete new edition of the entire book with lots of changes. This had significant changes to section 722 - both editorial and technical.
• Code of Practice (3rd edition). Aligns guidance to the new version of regulations.
• BS7671:2018 Amendment 1, published Feb 2020. This was published as a stand-alone amendment and had further changes to section 722 (only).

Concerning DC-sensitive RCDs, the requirements were:

• In the first CoP, "Type A RCDs are preferred, as they will provide some protection if the vehicle develops a DC fault to earth". No mention of Type B.
• In the first section 722 (introduced 2013, and identical in Amd3 2015): "The RCD protecting the charging point shall be at least a type A RCCB complying with BS EN 61008-1 or RCBO
  complying with BS EN 61009-1. If it is known that the d.c. component of the residual current exceeds 6 mA then a type B RCD complying with BS EN 62423 shall be installed."
• CoP 2nd edition just repeats the text from the regulation (above) and gives a description of what a Type A and Type B RCD are, but gives no extra guidance on which to use (apart from pointing out that Type AC is not permitted).
• BS7671:2018 has the 722.531 regulation on RCDs completely re-written. It still requires "at least Type A" for chargepoints in general, but then for chargepoints with Type1/Type2 connectors there's an additional requirement: "...protective measures against DC fault current shall be taken, except where provided by the EV charging equipment. The appropriate measures, for each connection point, shall be as follows: - RCD Type B; or -
  RCD Type A and appropriate equipment that provides disconnection of the supply in case of DC fault current above 6 mA."
• BS7671:2018Amd1 has again re-written this clause, but the requirements are essentially the same, just worded differently to take account of the fact that there are now standards for the 6mA DC detector devices: "(i) an RCD Type B, or (ii) an RCD Type A or Type F in conjunction with a residual direct current detecting device (RDC-DD) complying with BS IEC 62955 as appropriate to the nature of the residual and superimposed currents and
  recommendation of the manufacturer of the charging equipment."

So here there was a major change that came into force from Jan 2019: up until then you only had to provide Type B if you knew there was a need for it; afterwards you had to provide either Type B or Type A + DC detection device. Commando sockets (and 13A sockets) escape these new requrements - presumably on the assumption that the portable EVSE itself will contain the necessary RCD.


Concerning open PEN protection/use of PME earth:

• This has long been known as a problem for outdoor equipment, independent of any EV-specific issues. It is common practice to wire garden sheds and the like with a TT system.
• The first CoP advised that it's OK to use a PME earth inside a garage; for outdoors, to do risk assessment and use a separate TT earth for the chargepoint if there's no risk of touching the car and indoor equipment at the same time, otherwise to "consider" converting the whole house to TT.
• The original 722 said "a PME earthing facility shall not be used ... to charge a vehicle located outdoors" unless one of three requirements were satisfied; one related to 3-phase so irrelevant to most domestic installations, another referred to a device that nobody has ever made (different from the various devices that have appeared more recently), leaving just option (ii) of adding lots of extra earth electrodes in parallel with the PME earth. However, there was an additional statement "The requirements of this regulation need not be applied for the charging point at a dwelling, if none of (i), (ii) or (iii) is reasonably practicable." This left scope for argument over what is "reasonably practicable", which led some people to believe that you could freely use a PME earth in domestic installs since all of the other options were impracticable.
• 2nd edition CoP had considerable material to address the "reasonably practicable" and recommended that even if meeting the full requirements of the option (ii) approach is impractical you should still add as many extra earth rods as are practical rather than just giving up. It continued to recommend the use of the TT-island approach where risk assessment showed it was suitable.
• BS7671:2018 re-worded the 722.411 text, but still prohibiting the outdoor use of PME earth with the same three exceptions. The "reasonably practical" get-out was removed. The wording for the third option (open-PEN detection device) no longer refers to it needing to be connected to an earth electrode, so some people believe this change permits voltage-monitoring devices (though IMO it did not).
• BS7671:2018Amd1 keeps the same structure and the same options (i), (ii), (iii), but adds two new ones. (iv) provides the new option of a device providing open-PEN detection by monitoring the supply voltage, and option (v) allows proprietary approaches that achieve equal safety to (iii) or (iv).

So here the major change was this year when it became permissible to use the PME earth outdoors with a supply-voltage based open-PEN detector. Some people would argue that this approach was permitted from 2018, though that is highly questionable; prior to that it was definitely not permitted. Throughout the whole period and continuing today, TT-island (separate earth electrode for the chargepoint) has been permitted unless there's a problem with touching the car and indoor equipment at the same time. Use of the PME earth outdoors with no precaution was never permitted, unless you interpreted the "reasonably practical" get-out very broadly and ignore the guidance in the Code of Practice.

Note that OLEV-sponsored installs have always required compliance with the CoP and contemporary editions of BS7671 (as a condition of the grant). So if they are not complied with, that is a contractural matter for the homeowner, the installer and OLEV.

 

[Glan gluaisne]

The key part though is that it's always been known that there could well be a DC earth leakage component exceeding 6 mA, isn't it? The CP sits at +12V relative to earth when in state A, with a 1 k source impedance, so if shorted to earth there could be 12 mA of DC earth leakage.

 

[arg]

The key part though is that it's always been known that there could well be a DC earth leakage component exceeding 6 mA, isn't it? The CP sits at +12V relative to earth when in state A, with a 1 k source impedance, so if shorted to earth there could be 12 mA of DC earth leakage.

Maybe.

Short to earth isn't an issue - that's the normal current path and doesn't result in leakage. Short to line or neutral does result in leakage.

Short to line is hard to analyse; there would be significant (AC) current flowing, depending on the protection arrangements in the controller. With robust protection in the controller, the current is probably mostly AC and plenty to trip a 30mA RCD. With no protection in the controller, it probably blows up - if nothing more exciting happens, the 50W dissipated in that 1K resistor will probably burn it clear and disconnect; many other scenarios will result in the contactor opening either due to destruction of the controller or if the software is still alive, no longer detecting valid conditions on the pilot. It's hard to imagine a scenario with stable continued DC leakage.

Short to neutral is the one where sustained DC leakage is plausible. We're assuming PME, so N and E are connected back at the origin (and for a 30A circuit, that's going to be less than an ohm), so the short CP to E does drop the whole 12V across the 1K and give 12mA. However, this gives an invalid condition on the PE so the controller will open the contactor (required to be double-pole) and so remove the fault; if there's no car connected, nothing further happens (and no leakage); if the car is plugged in and wanting to charge, the thing will cycle at whatever retry rate the controller imposes, so in worst case you get pulsed DC leakage at that rate, not continuous. Alternatively, there could be a high-resistance short so that the CP remains in the valid range - 2.62V if the EVSE supports ventilation (vanishingly rare) or 5.58V for all normal EVSE. This gives a continuous DC leakage, but the car is loading with max 908R in this case, so there's a total of (12-5.58)/1000 A = 6.42mA flowing out of the controller but 5.68/908 A = 6.25mA flowing through the car and so only 0.2mA flowing as leakage.

The CP-neutral fault isn't hazardous in itself, so the only concern is the blinding of upstream RCDs by the DC leakage - a double-fault scenario. To get an entirely continuous DC requires a double-fault within the car charging system (CP-N short plus failure in the controller's logic/software) and then a third fault somewhere to actually cause a hazard. And this isn't a case of a potentially dormant first-fault allowing the second fault time to develop: the first fault will cause the chargepoint to stop working so the user is at least likely to stop using it (removing the fault) if not get it fixed. There is the case of the slowly pulsed DC with the controller retrying, but at worst that just extends disconnect times.

Finally, even if all of this does come to pass, how much of a problem is 12mA? Type-A RCDs are required to operate fully to spec with superimposed 6mA of pure DC - hence the 6mA spec for the DC fault monitoring device - but real-life devices will normally be better than that and work, perhaps with impaired performance, well over 6mA. So it is fairly likely that the RCD will still work when needed even in the presence of 12mA DC.


I admit that the Type-B or 6mA-protection is an improvement, but the level of improvement is fairly small. I think it is hard to justify more than a C3 on EICR for an installation that has Type-A RCD.

Of course there are plenty of installations around that were never compliant at the time they were installed - outdoor use of PME with no mitigation, Type-AC RCD etc, which do warrant a C2.

 

[Glan gluaisne]

Maybe.

Short to earth isn't an issue - that's the normal current path and doesn't result in leakage. Short to line or neutral does result in leakage.

Short to line is hard to analyse; there would be significant (AC) current flowing, depending on the protection arrangements in the controller. With robust protection in the controller, the current is probably mostly AC and plenty to trip a 30mA RCD. With no protection in the controller, it probably blows up - if nothing more exciting happens, the 50W dissipated in that 1K resistor will probably burn it clear and disconnect; many other scenarios will result in the contactor opening either due to destruction of the controller or if the software is still alive, no longer detecting valid conditions on the pilot. It's hard to imagine a scenario with stable continued DC leakage.

Short to neutral is the one where sustained DC leakage is plausible. We're assuming PME, so N and E are connected back at the origin (and for a 30A circuit, that's going to be less than an ohm), so the short CP to E does drop the whole 12V across the 1K and give 12mA. However, this gives an invalid condition on the PE so the controller will open the contactor (required to be double-pole) and so remove the fault; if there's no car connected, nothing further happens (and no leakage); if the car is plugged in and wanting to charge, the thing will cycle at whatever retry rate the controller imposes, so in worst case you get pulsed DC leakage at that rate, not continuous. Alternatively, there could be a high-resistance short so that the CP remains in the valid range - 2.62V if the EVSE supports ventilation (vanishingly rare) or 5.58V for all normal EVSE. This gives a continuous DC leakage, but the car is loading with max 908R in this case, so there's a total of (12-5.58)/1000 A = 6.42mA flowing out of the controller but 5.68/908 A = 6.25mA flowing through the car and so only 0.2mA flowing as leakage.

The CP-neutral fault isn't hazardous in itself, so the only concern is the blinding of upstream RCDs by the DC leakage - a double-fault scenario. To get an entirely continuous DC requires a double-fault within the car charging system (CP-N short plus failure in the controller's logic/software) and then a third fault somewhere to actually cause a hazard. And this isn't a case of a potentially dormant first-fault allowing the second fault time to develop: the first fault will cause the chargepoint to stop working so the user is at least likely to stop using it (removing the fault) if not get it fixed. There is the case of the slowly pulsed DC with the controller retrying, but at worst that just extends disconnect times.

Finally, even if all of this does come to pass, how much of a problem is 12mA? Type-A RCDs are required to operate fully to spec with superimposed 6mA of pure DC - hence the 6mA spec for the DC fault monitoring device - but real-life devices will normally be better than that and work, perhaps with impaired performance, well over 6mA. So it is fairly likely that the RCD will still work when needed even in the presence of 12mA DC.


I admit that the Type-B or 6mA-protection is an improvement, but the level of improvement is fairly small. I think it is hard to justify more than a C3 on EICR for an installation that has Type-A RCD.

Of course there are plenty of installations around that were never compliant at the time they were installed - outdoor use of PME with no mitigation, Type-AC RCD etc, which do warrant a C2.

I agree, the fault scenario that may cause RCD blinding with the potential for an electric shock isn't that likely, but the question relates to the framing of the words in the regs. We all know that the IET works in mysterious ways at times (metal CUs to fix poor terminal design and use, for example), but there was clearly a view that DC leakage greater than 6mA might be present for an EV charge point installation. I'm assuming that they considered that the cause of this was primarily the CP DC component, coupled with, perhaps, some DC leakage from charger in the car.

It is hard for any manufacturer to argue that this section from the old regs didn't apply to a charge point, IMHO: "If it is known that the d.c. component of the residual current exceeds 6 mA then a type B RCD complying with BS EN 62423 shall be installed." Presumably, Rolec et al might claim that they didn't know that there could be more than 6mA of DC residual current, which then implies that they don't properly understand how the CP signalling protocol works.

In the case of Rolec, this may have been the case, as all they seem to have done is fit off-the-shelf components into an enclosure, they don't seem to have had any input into the Viridian/Mainpine EVSE design, AFAIK, as that module (the original, Mk 1, without any built-in protection) was also used by other charge point manufacturers. I believe the original Mainpine EVSE predates even the CoP, and was primarily designed in accordance with the J1772 protocol, with it not intending to be able to provide any electrical safety protection capabilities outwith those in J1772.

I don't agree with marking this as a C3, although arguably it could be, given that we no longer have the option of using C4 (for those not familiar with the regs, there used to be an option to mark an installation as compliant with the regulations that applied when it was installed). Given the tendency in recent years for the various assessment and certification bodies to encourage electricians to tighten up markings on EICRs, so that things that are really non-issues, like dimensionally and electrically identical MCBs of a different make to the CU manufacturer now warranting a C2 (had one of these only a few weeks ago), then it seems reasonable to mark a Type A RCD/RCBO as C2 if it's protecting a charge point.

I don't wholly agree with some of the daftness that NICEIC, NAPIT etc are inflicting on everyone, something that seems to be getting worse through social media, but we are where we are, and it seems that they are intent on tightening things up, and not always in the best interest of balancing safety with value.

 

[arg]

It is hard for any manufacturer to argue that this section from the old regs didn't apply to a charge point, IMHO: "If it is known that the d.c. component of the residual current exceeds 6 mA then a type B RCD complying with BS EN 62423 shall be installed." Presumably, Rolec et al might claim that they didn't know that there could be more than 6mA of DC residual current, which then implies that they don't properly understand how the CP signalling protocol works.

I think we can both agree that the old wording was ambiguous. Possibly your reading of it was that intended by the authors, but if so hardly anybody read it that way - all of the big-name installers were just using Type-A, and the first two editions of the CoP give no further guidance in that direction (1st Ed doesn't even mention Type B). In fact CoP 2nd Ed, in the section on on-street installations, talks about using Type S upstream of multiple chargepoints each with their Type-A or Type-B RCD: that would be inappropriate if you believed there might be 12mA of DC leakage (which won't trip either a 30mA Type B nor a Type-A, yet could reasonably blind a Type S).

Just trying to read that requirement without any background "If it is known that the d.c. component of the residual current exceeds 6 mA then a type B RCD ... shall be installed.", it could be saying:

• If the load is such that there's a standing DC current over 6mA the you need Type-B (I don't think this was the intent, but it's a reasonable reading of the words).
• If the DC component exceeds 6mA under any plausible fault condition then you need Type-B (I think this was the intent, and you seem to agree).

And in either of those cases it's saying that if it is known this problem is likely to arise then you need the Type-B, but in the general case where you don't know then it's not required - you don't have to fit one just on the off-chance.

By the logic in my earlier post, I would argue that plausible faults relating to the CP don't give rise to >6mA pure DC leakage, so the simple presence of an EVSE with CP doesn't turn this into a "known that it exceeds 6mA" situation.

At the time, I assumed this clause was referring to possible unusual charger designs that were likely to generate DC leakage (either fault or normal operation). I did wonder if the Renault Zoe, with its unusual non-isolated charger, was thought to be a case in point and so you might need Type-B for chargepoints installed for the use of such vehicles; however, Renault's instructions at the time didn't seem to require it.

 

[Glan gluaisne]

My view is that it's now hard to be unbiased, given that more evidence of the possible risk from net DC leakage currents seems to have been found in recent years, or at least, the level of awareness of the risk has significantly increased. The advent of more and more switched mode supply driven appliances, that may well produce supply artefacts that could blind old-style type AC, perhaps also type A, RCDs seems to have raised the level of awareness of the risk.

The same argument could be applied equally to the protection required for things like PV inverters, yet we know that the majority of PV systems installed for years had no effective protection against DC earth leakage, at best they may have had a type A RCD, but I wouldn't mind betting that some installations only had type AC protection. Like some newer charge points, some newer solar inverters have built in DC tolerant earth leakage protection, but there are thousands around that don't, despite there being a requirement for it. Clearly the electric shock risk is a bit lower, but not for the people offering solar panel cleaning services, perhaps.

 

[arg]

I certainly agree that with the vast majority of appliances having switching power supplies in them, pure 50Hz AC fault currents can no longer be assumed. Given the low incremental cost of Type-A, there's a strong argument for retiring Type-AC altogether. It's very easy to imagine fault scenarios giving rise to the type of pulsed-DC that Type-A will handle, but I haven't seen any good analysis of the probability of pure-DC leakage in common circuit configurations. I'm willing to believe the extra cost of Type-B is worth it for EV chargers, but more for the possibility of internal faults than the short-on-the-CP argument.

 

[bitmanEV]

I have indeed installed Type-B on Switched type PV inverters (transformer less) and Type AC on transformer PV Inverters like older the SMA SB
For my own use I have SMA Sunny Boys transformer PV Inverters due the fact less EMC as being a Radio Amateur and I can use my HF transceivers very close to my PV installation without any noise

 

[Glan gluaisne]

Those older Sunny Boys were a solid and reliable bit of kit. Mind you, they were a handful to fit high on a wall!

 

[MrBadger]

Moderator comment - this thread was spun off from Rolec EV Wallpod

Other threads of note which may or may not be merged in future are:

Tesla Wall Connector - Type B / Type A-EV RCD

 

[ACarneiro]

Honestly, seeing @arg and @Glan gluaisne bouncing off each other made me think Christmas has come early.
Few things are more fascinating that seeing two cogent, knowledgeable people arguing their points in a respectful and truly interested way.
Doesn’t happen nearly often enough any more these days.

My thanks to the both of you

 

[LukeUK]

Yup, thanks both. This has been informative, educational and overall very balanced. I appreciate reading such well put together and considered content. Thanks for your time and input.

Luke

 

[HKP_Tom]

Best thread this week. Useful, even if I don’t understand it all!

 

[Marvin03]

For anyone in Leeds I can thoroughly recommend Graham Scott at OLEV Approved EV Installations. I spoke with most of the local installers and some never showed up for surveys, others had massive waits, and many big outfits just wanted to do it there way with some armoured cable around the outside of the property from the meter box.

Installation was within a week of making contact Graham. He worked to my requirements and was more than happy with installing the specific Garo Type B + PEN fault unit that I wanted (I'd seen a lot of Tesla connectors installed with Type A RCDs, both on here and YouTube). Also, there was an option of fitting the wiring internally and with no visible cables which was great. Got all the installation certificate paperwork through within a few days.

I’ve an EO mini pro 2 being installed which will require external cable straight to meter box (according to fitter). However following recent building work I have 6mm internal cable fitted to new consumer unit with spare RCD slots. Is it an option to wire into a consumer unit?

 

[Glan gluaisne]

I’ve an EO mini pro 2 being installed which will require external cable straight to meter box (according to fitter). However following recent building work I have 6mm internal cable fitted to new consumer unit with spare RCD slots. Is it an option to wire into a consumer unit?

Possibly, but 6mm² is way too light for tails to a CU, the normal size would be 25mm², unless the main fuse was something like 60 A, when it would be OK to use 16mm² (although as the real difference in price between the two is tiny, it doesn't make sense to use 16mm², IMHO). Is this a second CU, fed from a protected way on a main CU? The supply cable to the charge point would normally be 6mm², unless it's very long, although in terms of current drawn, 6mm² is a bit of an overkill, it's usually specified both to reduce the voltage drop and because the manufacturers of charge points often stipulate 6mm² cable, perhaps because they've type approved the units and terminations with this size cable.

You can connect a charge point to a spare slot in a CU, as long it has the required protection built-in, although it's not ideal, IMHO, as drawing 32 A for hours on end tends to make MCBs/RCBOs run a bit warm, so if doing this it's a good idea to leave a space either side to allow heat to dissipate a bit more readily. Depending on the specific charge point, it may need to be connected to a non-RCD protected slot, with just over-current protection, or it may be able to be connected to a slot protected by an RCD.

If you choose to fit a charge point that doesn't have integral DC tolerant earth leakage protection, then you still need to fit a small CU to house a Type B or Type EV RCD, plus, perhaps, an MCB for over-currrent protection if this small CU is fed directly from the incoming supply.

If the house has a PME earthing system, then you also need to ensure that there is some form of open PEN protection, either built-in to the charge point, by using one of the open PEN units (which can also house an MCB and Type B/EV RCD) or by using a earth electrode.

 

[Marvin03]

Possibly, but 6mm² is way too light for tails to a CU, the normal size would be 25mm², unless the main fuse was something like 60 A, when it would be OK to use 16mm² (although as the real difference in price between the two is tiny, it doesn't make sense to use 16mm², IMHO). Is this a second CU, fed from a protected way on a main CU? The supply cable to the charge point would normally be 6mm², unless it's very long, although in terms of current drawn, 6mm² is a bit of an overkill, it's usually specified both to reduce the voltage drop and because the manufacturers of charge points often stipulate 6mm² cable, perhaps because they've type approved the units and terminations with this size cable.

You can connect a charge point to a spare slot in a CU, as long it has the required protection built-in, although it's not ideal, IMHO, as drawing 32 A for hours on end tends to make MCBs/RCBOs run a bit warm, so if doing this it's a good idea to leave a space either side to allow heat to dissipate a bit more readily. Depending on the specific charge point, it may need to be connected to a non-RCD protected slot, with just over-current protection, or it may be able to be connected to a slot protected by an RCD.

If you choose to fit a charge point that doesn't have integral DC tolerant earth leakage protection, then you still need to fit a small CU to house a Type B or Type EV RCD, plus, perhaps, an MCB for over-currrent protection if this small CU is fed directly from the incoming supply.

If the house has a PME earthing system, then you also need to ensure that there is some form of open PEN protection, either built-in to the charge point, by using one of the open PEN units (which can also house an MCB and Type B/EV RCD) or by using a earth electrode.

Thx. Everything is brand new - meter and box (100A of course) and new CU with 50A RCD but that’s about the extent of my electrical knowledge. Cable run internally from CU to exit point in wall for charger is approx 5m. Similar for external run if that is what is needed - the meter box is on the wall externally and CU other side internally.

sounds like I should let the installer do their thing and install direct to meter and external armoured cable.

 

[Glan gluaisne]

It's better to have the charge point fed from a small additional CU fed directly from the incoming supply if it's practical to do this, IMHO, really because the charge point could be running at 32 A for several hours, so the MCB/RCBO feeding it will warm up, and heat build up within CUs can be a slight problem.

In your case, the EO Mini doesn't have built-in open PEN protection, so if your house has a PME earthing arrangement you need an open PEN device fitted to the supply, and this comes in a small enclosure that needs to be fitted in the supply to the charge point anyway. The EO Mini also needs it's own Type A RCD (it has built in DC tolerant protection), plus the cable needs over current protection, and this can be included in the same enclosure as the open PEN device.

 

[Morriscoombes]

I insisted the install of my Zappi2 was to a small extra CU for that reason. Clearly delineated from rest of system.