Ok, so I'm still a bit confused about the characteristics of the various types of RCDs, i.e. AC/A/F/B. I've concluded that's because there are two classes of effects: a type of residual current that will cause a trip, and a type of residual current that will blind the device against other types of current it would normally trip on. I find that online discussions and 531.3.3 don't clearly distinguish these cases. Can anyone provide, or give a link to, a clear description of what each type RCD is designed to trip on, and what can blind it? For example, might a smooth DC residual current >6ma blind a type A? And if so, is that completely blind, or just blind against pulsating DC (AC still works)? Etc.
It is indeed a mess. I reckon in about 10 years all new RCDs will have DC sensing.
Right now mass production is mainly making AC and A types as this is an established design, and a small variation on it.
Blinding requires an AC-DC conversion on the mains side of any transformers, and a fault to earth from the DC bus.
Historically it has not been much of a problem, but high power class one devices that are switch mode like VSD and EV chargers
are a new problem. Here we do have the mains being rectified and smoothed as the first step, and also a circuit that then slices that DC bus at a supersonic frequency, which is after all how we keep the transformers so small and light.
But that brings 2 new problems.
1) there is now a credible DC source in an earthed metal box (smaller switching supplies tend to be double insulated so the earth fault is impossible)
2) to stop the switching frequency radiating off the wiring like a crude radio transmitter, we now need to add high frequency filtering, so there are now RF capacitors between line and earth, among other things, so there is a standing leakage to earth even when all is well.
So the RCD designer is being squeezed, as now they must neither trip nor blind with non-mains frequencies, they must not blind on DC faults, and must work with a significant leakage and not trip.
The price and scarcity of devices to the harder specs tells us this is a hard problem, and that situation will remain until the 2nd tier manufacturers can get tooled up for it.
It is indeed a mess. I reckon in about 10 years all new RCDs will have DC sensing.
Right now mass production is mainly making AC and A types as this is an established design, and a small variation on it.
Blinding requires an AC-DC conversion on the mains side of any transformers, and a fault to earth from the DC bus.
Historically it has not been much of a problem, but high power class one devices that are switch mode like VSD and EV chargers
are a new problem. Here we do have the mains being rectified and smoothed as the first step, and also a circuit that then slices that DC bus at a supersonic frequency, which is after all how we keep the transformers so small and light.
But that brings 2 new problems.
1) there is now a credible DC source in an earthed metal box (smaller switching supplies tend to be double insulated so the earth fault is impossible)
2) to stop the switching frequency radiating off the wiring like a crude radio transmitter, we now need to add high frequency filtering, so there are now RF capacitors between line and earth, among other things, so there is a standing leakage to earth even when all is well.
So the RCD designer is being squeezed, as now they must neither trip nor blind with non-mains frequencies, they must not blind on DC faults, and must work with a significant leakage and not trip.
The price and scarcity of devices to the harder specs tells us this is a hard problem, and that situation will remain until the 2nd tier manufacturers can get tooled up for it.
