Regulation stating a type AC RCD can not be upstream from a type A RCD

Further to Dave's I'm not convinced that there is a credible  L-E (or N-E  but not both at once ) fault path that puts un-smoothed DC into the loop (which all that an A type is sure to detect) . Certainly it is nothing  like AC leakage that occurs by design in common mode filters.

The car chargers do use the CPC in the link from car to charger to carry a control signal that is uni-polar - not really a continuous DC like a battery but certainly a DC of sorts.. But even though this flows in the CPC, how exactly that would saturate an L-N measuring RCD is not at all obvious. Certainly in normal operations it won't and a credible fault where it does is likely to involve enough fault current to trip everything  else as well.

In many ways things like an LE fault on the lamp side of the traditional filament lamp dimmer switch are more of a hazard, as the series triac can fail in  way that makes it fire assymetrically and the light fitting may be class 1 , There  are a few appliances like cheap hair-dryers that use a series diode for the hi/lo heat setting, but these are all double insulated and have been since about 1970, and before that rectifiers would have been more expensive than creating  a heater coil with a centre tap contact. (time  was when 5s was the cost of a single diode). But we are still looking at double fault to danger at least to get that 'blinding' fault path.

Also  even in odd cases when it occurs 'blinding' is not all or nothing affair, more of a de-sensitising, usually more or less pro-rata, so to blind a 30mA RCD to a 100mA fault, something like 100mA of DC is needed, as with much less the core is pulled out of saturation on the  half cycles that oppose the DC bias, and an AC is detected in the trip side. Not great, but  you do need a lot of unsmoothed DC to make the situation dangerous.

I think the risk is misunderstood and often overstated.

Mike.

Further to Dave's I'm not convinced that there is a credible  L-E (or N-E  but not both at once ) fault path that puts un-smoothed DC into the loop (which all that an A type is sure to detect) . Certainly it is nothing  like AC leakage that occurs by design in common mode filters.

The car chargers do use the CPC in the link from car to charger to carry a control signal that is uni-polar - not really a continuous DC like a battery but certainly a DC of sorts.. But even though this flows in the CPC, how exactly that would saturate an L-N measuring RCD is not at all obvious. Certainly in normal operations it won't and a credible fault where it does is likely to involve enough fault current to trip everything  else as well.

In many ways things like an LE fault on the lamp side of the traditional filament lamp dimmer switch are more of a hazard, as the series triac can fail in  way that makes it fire assymetrically and the light fitting may be class 1 , There  are a few appliances like cheap hair-dryers that use a series diode for the hi/lo heat setting, but these are all double insulated and have been since about 1970, and before that rectifiers would have been more expensive than creating  a heater coil with a centre tap contact. (time  was when 5s was the cost of a single diode). But we are still looking at double fault to danger at least to get that 'blinding' fault path.

Also  even in odd cases when it occurs 'blinding' is not all or nothing affair, more of a de-sensitising, usually more or less pro-rata, so to blind a 30mA RCD to a 100mA fault, something like 100mA of DC is needed, as with much less the core is pulled out of saturation on the  half cycles that oppose the DC bias, and an AC is detected in the trip side. Not great, but  you do need a lot of unsmoothed DC to make the situation dangerous.

I think the risk is misunderstood and often overstated.

Mike.

mapj1 said:

The car chargers do use the CPC in the link from car to charger to carry a control signal that is uni-polar - not really a continuous DC like a battery but certainly a DC of sorts.. But even though this flows in the CPC, how exactly that would saturate an L-N measuring RCD is not at all obvious.

It's not obvious, but with an N-PE fault between EVSE and vehicle, there's a path to share the DC steady-state or pulsed current back to PE through the supply N-PE link (in PME installations, at the service head, so we are talking similar impedances and sharing is a definite mechanism for this to occur).

As David says, not all Type AC RCDs can be saturated, but some can, and this can be demonstrated easily.

Thanks, one I had missed,  so with that in mind, to recover double fault to danger protection then  for the lead out to the car, we need to insulate and sheath the CPC to the same standard as if it was  another  live core.

Is that not already the case ?

Mike.

In the cable, yes, but in the vehicle and EVSE (downstream of the pilot circuit), you may be back to insulated N, and exposed-conductive-part as PE?