DC on AC supply

Weirdbeard:

Nathaniel:

Weirdbeard:

What I have wondered with regards this situation is if an installation has equipment that can apparently disable commonly found RCDs, would this disablement also apply to other installations that share a supply, for example with looped supplies? 

This is a question I got elsewhere a year ago, mainly about worries with solar PV installations and EV chargers affecting other customers on the network by their possible dc, and so a need to have type-B RCDs as default regardless of whether the installation included such devices. The answer: practically, no. 

 

Thanks for the reply, are you aware of any common consumer appliances that are likely to defeat a typical 30mA RCD and render the RCD useless as additional protection against electric shock?

The answer: no. 

I don't usually manage such short answers, as there are so many ifs and buts in this sort of subject. But you helped me by including the word 'aware', and then further helped with 'typical' and 'useless'.  And - please note - I'm assuming that the appliance is in its intended state, not having a fault.

aware:

I've had several reasons to study RCD standards and actual behaviour, particularly with passive RCDs and dc, so I'm quite confident about that situation. But I haven't done so much with high-frequency disturbance or different types of voltage-dependent RCD. Many modern RCBOs are voltage-dependent in order to make them easily fit in 1 module.

For appliances and inverters I've not looked thoroughly at what extremes the product standards would permit for protective-conductor currents at dc and different frequencies. Also, if each device is permitted some small level of dc in the PE, then when dozens are connected and by bad luck the currents mainly sum in the same direction, they could anyway exceed a threshold like 6 mA. In the case I mentioned earlier, the question I received assumed a worst case of 6 mA dc from each inverter or EV, so it was enough to show that there wasn't a practical problem (for other customers) in their type of system even with many times this current.

I've checked PE-currents for only a few dozen varied household electronic loads (IT equipment, induction hobs, heatpumps, washer/dryer), and found only low dc current (<1 mA but without good measurement in this range) in the PE, and at most about 10 mA rms 'noise' at higher frequencies during some parts of a wash cycle. The L-N dc wasn't considered: RCDs can have a little influence from L-N currents when these aren't well balanced in the way the wires are wrapped, but they're tested to not trip at 6*In. So 96 A in a 16 A device shouldn't give an imbalance as much as the residual trip level of e.g. 15 mA to 30 mA for a 30 mA RCD. Taking the same ratio for dc, tens of amps L-N dc would be needed in order to get 6 mA imbalance.

typical:

The standards for RCDs are based on passing a few tests. RCDs tend to be made in quite similar ways governed by familiarity, physics and economics. Performance outside the specified tests can then be quite well guessed within some range. But if we go by the letter of the standard, then a type-A wouldn't (from what I remember of IEC61008) even have to work with superimposed 6 mA dc and full-wave 30 mA ac, as the dc test is only done with half-wave ac. So situations that are typically safe may differ from situations that must be safe within the requirements of a standard.  A month or two ago I linked to the Australian state that requires a further test not included in the IEC standard, to check that voltage-dependent RCBOs wouldn't simply burn out their tripping coil on the first operation in the event that they're fed with the load terminals connected to the line (source).  Weak connection to VW and diesels, regarding standards and reality.

useless:

if the RCD can still work for some possible waveforms of the fault that needs additional protection, or its threshold is only mildly increased, then I'd argue it's not useless, but just not so useful...

 

Weirdbeard:

Nathaniel:

Weirdbeard:

What I have wondered with regards this situation is if an installation has equipment that can apparently disable commonly found RCDs, would this disablement also apply to other installations that share a supply, for example with looped supplies? 

This is a question I got elsewhere a year ago, mainly about worries with solar PV installations and EV chargers affecting other customers on the network by their possible dc, and so a need to have type-B RCDs as default regardless of whether the installation included such devices. The answer: practically, no. 

 

Thanks for the reply, are you aware of any common consumer appliances that are likely to defeat a typical 30mA RCD and render the RCD useless as additional protection against electric shock?

The answer: no. 

I don't usually manage such short answers, as there are so many ifs and buts in this sort of subject. But you helped me by including the word 'aware', and then further helped with 'typical' and 'useless'.  And - please note - I'm assuming that the appliance is in its intended state, not having a fault.

aware:

I've had several reasons to study RCD standards and actual behaviour, particularly with passive RCDs and dc, so I'm quite confident about that situation. But I haven't done so much with high-frequency disturbance or different types of voltage-dependent RCD. Many modern RCBOs are voltage-dependent in order to make them easily fit in 1 module.

For appliances and inverters I've not looked thoroughly at what extremes the product standards would permit for protective-conductor currents at dc and different frequencies. Also, if each device is permitted some small level of dc in the PE, then when dozens are connected and by bad luck the currents mainly sum in the same direction, they could anyway exceed a threshold like 6 mA. In the case I mentioned earlier, the question I received assumed a worst case of 6 mA dc from each inverter or EV, so it was enough to show that there wasn't a practical problem (for other customers) in their type of system even with many times this current.

I've checked PE-currents for only a few dozen varied household electronic loads (IT equipment, induction hobs, heatpumps, washer/dryer), and found only low dc current (<1 mA but without good measurement in this range) in the PE, and at most about 10 mA rms 'noise' at higher frequencies during some parts of a wash cycle. The L-N dc wasn't considered: RCDs can have a little influence from L-N currents when these aren't well balanced in the way the wires are wrapped, but they're tested to not trip at 6*In. So 96 A in a 16 A device shouldn't give an imbalance as much as the residual trip level of e.g. 15 mA to 30 mA for a 30 mA RCD. Taking the same ratio for dc, tens of amps L-N dc would be needed in order to get 6 mA imbalance.

typical:

The standards for RCDs are based on passing a few tests. RCDs tend to be made in quite similar ways governed by familiarity, physics and economics. Performance outside the specified tests can then be quite well guessed within some range. But if we go by the letter of the standard, then a type-A wouldn't (from what I remember of IEC61008) even have to work with superimposed 6 mA dc and full-wave 30 mA ac, as the dc test is only done with half-wave ac. So situations that are typically safe may differ from situations that must be safe within the requirements of a standard.  A month or two ago I linked to the Australian state that requires a further test not included in the IEC standard, to check that voltage-dependent RCBOs wouldn't simply burn out their tripping coil on the first operation in the event that they're fed with the load terminals connected to the line (source).  Weak connection to VW and diesels, regarding standards and reality.

useless:

if the RCD can still work for some possible waveforms of the fault that needs additional protection, or its threshold is only mildly increased, then I'd argue it's not useless, but just not so useful...