Double wound safety transformer for EV supply.

Interestingly no one has suggested a reason why a charging RCD should be DC sensitive.

From what I've been able to gather, it's not safe (at present) to assume that there's any galvanic isolation between the AC mains supply and the vehicle's propulsion battery & associated circuitry when on change - indeed it seems that at least one current model re-uses the charger power electronics in the power train (I'm guessing during regenerative breaking) so it's all quite tightly coupled together. It seems they don't use the traditional chassis return for the d.c. side (at least not for the propulsion system) but as usual with vehicle electrics there's only single insulation between the internal wiring and chassis. While on charge the vehicle's chassis/bodywork is likely to be connected to the supply's protective conductor. So, as one example, a single fault from say the battery circuit (at perhaps a few hundred volts d.c.) to the chassis, while on charge from a conventional TN or TT supply, would potentially mean a d.c. fault current flow from the battery, to bodywork, via the protective conductor to the EVSE and hence via the means of earthing back to the supply N point and then along the N back through the vehicle's charging circuit to the battery -ve. That d.c. fault current would be flowing through the N side of any RCDs within that loop - both on the EVSE circuit and upstream which may also protect other circuits - potentially blinding them (D-loc style) to any imbalance. Similarly faults from within the charging circuit or elsewhere in the vehicles' wiring could mean pulsed rather than smooth d.c. or at a whole variety of driving voltages.

 

An interesting reply Andy, but the earth connection to the car is not a protective conductor in the sense of a conventional circuit, it is simply there to enable the electronics to power the car charger up. If you think about it, it doesn't provide any protection to anything, as the supply is IT, and a double fault etc. is required to get any dangerous situation.

We might only have in mind the idea of defeating the vehicle's earth monitoring electronics, but as I see it the protective conductor between the EVSE and vehicle does a lot more than that. Both the vehicle and the EVSE potentially have exposed-conductive-parts and we can easily end up with a hazardous situation if they're not connected together in the suggested system - for example if we didn't have the protective conductor between EVSE and vehicle and the EVSE's protective conductor was connected to the separated circuit's "N" as suggested then a single L-chassis fault in the vehicle would immediately put 230V between the two parts, which are quite likely within reach of each other.


The system as claimed isn't a IT system, but a separated one - they are slightly but significantly different (not least because exposed-conductive-parts are deliberately referenced to earth in an IT system, and if it were an IT system BS 7671 would insist on some kind of insulation monitoring).

It would ensure that the RCD tripped if the live faulted to the car bodywork, but even so this would not be a danger to any person or livestock. If there was a neutral fault the RCD would probably trip due to current diverted via the fault, but again could in no way be considered dangerous. The "even so" means that the RCD is not the primary protection of the circuit, it is additional protection against multiple faults, single ones not being dangerous.

But with the suggested system we've in effect built-in the first fault by connecting exposed-conductive-parts (via the EVSE's c.p.c) to a separated live conductor (let's call it "N" for the sake of argument). Our first fault (which would be a 2nd fault in a conventional separated system) now causes a large fault current to flow in the c.p.c.s and given the impedance of the protective conductors is likely to be a significant fraction of the overall fault loop impedance, create reasonable potential difference between different exposed-conductive-parts (the EVSE and vehicle for example) - so rapid disconnection (e.g. within the times of table 41.1) is then necessary to avoid danger (which is exactly what 418.3.7 tells is we need to do for 2nd faults in the general case of separated circuits). So the RCD isn't providing some additional protection, it's providing basic ADS.


I've got in mind the video that Zoomup posted a few days ago showing the installation of EVSE (somewhere abroad) - where the "pod" clamped to a steel pole which ran through the main wiring area and was shown (quite reasonably) being earthed to the supply c.p.c. as an exposed-conductive-part. That pole was then sunk into the ground to support the EVSE - hence would be picking up an earth potential in just the same way as an extraneous-conductive-part (even if may miss being defined as such by being part of the electrical installation). If we applied the CoP's approach of supplying this though an isolating transformer but with the EVSE's c.p.c. connected to the separated circuit's "N" then what we've actually created is nothing less than a simple TN-S supply - with none of the advantages of a separated system at all (but with a lot of extra expense, both in initial costs an on-going electrical losses). We might as well just have created a TT island.


Some of the problems could of course be mitigated if the EVSE had no exposed-conductive-parts, but that doesn't seem to be a safe assumption, especially for EVSE in public/highway areas where substantial steel enclosures seem to be the norm for mitigating against both vandalism and impact.

 

It has been extremely difficult to design a completely satisfactory and safe car charging installation with a PME supply, and particularly with the possibility of a broken CNE. This scheme is probably as safe as is possible when cars are class 1, and I would certainly recommend that all new car designs are class 2, because that change is fairly easy and cheap, and removes the charging dangers unless damaged flexible cables are involved.

Now there I can agree completely! (Although I anticipate some objection due to the weight of the on-board isolating transformer that would probably be needed)


   - Andy.

Interestingly no one has suggested a reason why a charging RCD should be DC sensitive.

From what I've been able to gather, it's not safe (at present) to assume that there's any galvanic isolation between the AC mains supply and the vehicle's propulsion battery & associated circuitry when on change - indeed it seems that at least one current model re-uses the charger power electronics in the power train (I'm guessing during regenerative breaking) so it's all quite tightly coupled together. It seems they don't use the traditional chassis return for the d.c. side (at least not for the propulsion system) but as usual with vehicle electrics there's only single insulation between the internal wiring and chassis. While on charge the vehicle's chassis/bodywork is likely to be connected to the supply's protective conductor. So, as one example, a single fault from say the battery circuit (at perhaps a few hundred volts d.c.) to the chassis, while on charge from a conventional TN or TT supply, would potentially mean a d.c. fault current flow from the battery, to bodywork, via the protective conductor to the EVSE and hence via the means of earthing back to the supply N point and then along the N back through the vehicle's charging circuit to the battery -ve. That d.c. fault current would be flowing through the N side of any RCDs within that loop - both on the EVSE circuit and upstream which may also protect other circuits - potentially blinding them (D-loc style) to any imbalance. Similarly faults from within the charging circuit or elsewhere in the vehicles' wiring could mean pulsed rather than smooth d.c. or at a whole variety of driving voltages.

 

An interesting reply Andy, but the earth connection to the car is not a protective conductor in the sense of a conventional circuit, it is simply there to enable the electronics to power the car charger up. If you think about it, it doesn't provide any protection to anything, as the supply is IT, and a double fault etc. is required to get any dangerous situation.

We might only have in mind the idea of defeating the vehicle's earth monitoring electronics, but as I see it the protective conductor between the EVSE and vehicle does a lot more than that. Both the vehicle and the EVSE potentially have exposed-conductive-parts and we can easily end up with a hazardous situation if they're not connected together in the suggested system - for example if we didn't have the protective conductor between EVSE and vehicle and the EVSE's protective conductor was connected to the separated circuit's "N" as suggested then a single L-chassis fault in the vehicle would immediately put 230V between the two parts, which are quite likely within reach of each other.


The system as claimed isn't a IT system, but a separated one - they are slightly but significantly different (not least because exposed-conductive-parts are deliberately referenced to earth in an IT system, and if it were an IT system BS 7671 would insist on some kind of insulation monitoring).

It would ensure that the RCD tripped if the live faulted to the car bodywork, but even so this would not be a danger to any person or livestock. If there was a neutral fault the RCD would probably trip due to current diverted via the fault, but again could in no way be considered dangerous. The "even so" means that the RCD is not the primary protection of the circuit, it is additional protection against multiple faults, single ones not being dangerous.

But with the suggested system we've in effect built-in the first fault by connecting exposed-conductive-parts (via the EVSE's c.p.c) to a separated live conductor (let's call it "N" for the sake of argument). Our first fault (which would be a 2nd fault in a conventional separated system) now causes a large fault current to flow in the c.p.c.s and given the impedance of the protective conductors is likely to be a significant fraction of the overall fault loop impedance, create reasonable potential difference between different exposed-conductive-parts (the EVSE and vehicle for example) - so rapid disconnection (e.g. within the times of table 41.1) is then necessary to avoid danger (which is exactly what 418.3.7 tells is we need to do for 2nd faults in the general case of separated circuits). So the RCD isn't providing some additional protection, it's providing basic ADS.


I've got in mind the video that Zoomup posted a few days ago showing the installation of EVSE (somewhere abroad) - where the "pod" clamped to a steel pole which ran through the main wiring area and was shown (quite reasonably) being earthed to the supply c.p.c. as an exposed-conductive-part. That pole was then sunk into the ground to support the EVSE - hence would be picking up an earth potential in just the same way as an extraneous-conductive-part (even if may miss being defined as such by being part of the electrical installation). If we applied the CoP's approach of supplying this though an isolating transformer but with the EVSE's c.p.c. connected to the separated circuit's "N" then what we've actually created is nothing less than a simple TN-S supply - with none of the advantages of a separated system at all (but with a lot of extra expense, both in initial costs an on-going electrical losses). We might as well just have created a TT island.


Some of the problems could of course be mitigated if the EVSE had no exposed-conductive-parts, but that doesn't seem to be a safe assumption, especially for EVSE in public/highway areas where substantial steel enclosures seem to be the norm for mitigating against both vandalism and impact.

 

It has been extremely difficult to design a completely satisfactory and safe car charging installation with a PME supply, and particularly with the possibility of a broken CNE. This scheme is probably as safe as is possible when cars are class 1, and I would certainly recommend that all new car designs are class 2, because that change is fairly easy and cheap, and removes the charging dangers unless damaged flexible cables are involved.

Now there I can agree completely! (Although I anticipate some objection due to the weight of the on-board isolating transformer that would probably be needed)


   - Andy.