Can anyone explain type A RCD's

[oscar21]

I get how they work and can kind of understand the reasoning, I remember the early "no-trip" RCD testers, D-LOG made one that saturated the RCD core with DC to stop it tripping whilst doing an earth loop test so I agree that stray DC leakeage could stop an RCD tripping.

But, lets say you install a car charger you would protect it with a type A RCD/RCBO but would you also change the main RCD's in the consumer unit if they were AC type because if the car charger does leak DC then it could interfere with the house ones as well. And if it interferes with the household ones, what about next door's RCD's it could also stop them from working couldn't it. how does the DC know to stop propagating down the cables and stopping things from working? Or is it just a pointless reg that has bee brought upon us to keep the coin rolling in.

 

[pc1966]

You need a lot of DC to kill off RCDs. The difference between type AC and type A is not much really, just the saturation point of the sense transformer (and so what is guaranteed). type B is another matter, it has to use something like Hall effect or flux-gate technology to measure pure DC.

What happens with your neighbour is not important in virtually all cases, it is the nature of the fault on your RCD that determines how much DC is present and if that can suppress the AC aspect that is usually what will kill.

See also: RCD "blinding" by DC - https://www.electriciansforums.net/threads/rcd-blinding-by-dc.193857/

 

[loz2754]

You're correct in assuming that any upstream RCDs would need to be at least type A as well. The BEAMA guide to RCDs has some good info, along with diagrams explaining how to achieve selectivity with RCDs in series.

 

[Lucien Nunes]

what about next door's RCD's it could also stop them from working couldn't it. how does the DC know to stop propagating down the cables

DC only matters to an RCD if it is differential i.e. an imbalance.

DC leakage (and indeed DC components of distorted AC load waveforms, which are likely to be much larger in magnitude) causes DC to flow in the mains. Most of it passes through the supply transformer as that offers the lowest impedance, but some can end up passing through other parallel-connected loads. E.g. the mains transformer in a large home cinema receiver was found to be humming excessively because of DC superimposed on the mains by a nearby rogue PSU.

However as far as any other RCDs are concerned, even if there is unwanted DC passing through the loads they serve, it's balanced current not differential leakage, so it doesn't contribute to the core flux and doesn't blind them.

 

[Rockingit]

Or is it just a pointless reg that has bee brought upon us to keep the coin rolling in.

Often we'd say yes, but having personally witnessed a continuous cumulative 7A flowing to earth via type AC devices that didn't blink due to DC blinding, I'd say the powers at be have a point on this one.

 

[nicebutdim]

Often we'd say yes, but having personally witnessed a continuous cumulative 7A flowing to earth via type AC devices that didn't blink due to DC blinding, I'd say the powers at be have a point on this one.

I'd further contend that this issue was driven by self-regulation within the industry and it seemed to me that the IET were far behind this particular curve.

 

[littlespark]

Hypothetical, but likely scenario.

Older split CU with type AC RCDs.
New EV added to system but in parallel to existing, not fed from.

New EV having its own type A.

Does the main board RCDs need changed? Considering they are in parallel, not upstream.

I would think anything next door would be too far away to be affected, but within a metre in the same property?

 

[pc1966]

I'd further contend that this issue was driven by self-regulation within the industry and it seemed to me that the IET were far behind this particular curve.

Indeed, in the EU type AC went out a long time ago.

Part of that could be down to widespread TT use (so they are more dependent on RCD for fault disconnection) but equally there should be very little cost difference between AC and A as it is just a higher saturation point core for the differential transformer. Practically nothing else needs to change.

Type B, as above, is a whole different kettle of fish...

 

[pc1966]

Older split CU with type AC RCDs.
New EV added to system but in parallel to existing, not fed from.

New EV having its own type A.

Does the main board RCDs need changed? Considering they are in parallel, not upstream.

Not required for the EV as it is on separate RCD.

Maybe worth considering if they have lots of electronics and its a TT setup, but I would not be to worried for indoor TN stuff.

 

[Lucien Nunes]

Does the main board RCDs need changed? Considering they are in parallel, not upstream.
I would think anything next door would be too far away to be affected, but within a metre in the same property?

We can discover how the DC propagates through the upstream / parallel-connected system by comparing impedances. In the case you mention, we are interested in the impedance of the supply (that feeds both the main and additional RV boards) versus the IR of the circuits connected to the main board RCDs, to calculate the fraction of any DC leakage from the EV board that could stray via the main board RCD and blind it. Note that it is the IR of the main board circuits, not the resistance of the loads themselves connected to those circuits, because we're only interested in unbalanced currents. DC circulating through the loads might not be desirable, but it doesn't blind the RCD so it's not relevant to us here.

Taking some realistic impedances for comparison, the supply impedance feeding both the main and additional EV boards might be 0.1 ohms, while the lowest L-E IR on a faulty circuit on the main board that would still hold on a 30mA device (at 20mA) would be around 11.5kΩ. If, then the EV circuit started to leak 10mA DC, that would divide in the ratio 115000:1, with just 0.09μA passing through the main board RCD as imbalance and the remainder returning to the supply. So there won't be any measurable influence on the RCD performance.

Suppose instead there was an N-E fault on the main board circuit. On a TN-S supply there might be half a dozen volts present N-E and the minimum N-E IR that will hold might be taken as 6 / 0.02 = 300Ω. Now we have a ratio 3000:1 and something like 3.3μA DC will pass through the main board RCD, still a trivial current.

In a nutshell, unless there is some characteristic of the supply that drastically raises its effective resistance to DC, almost all of the leakage will pass directly upstream and not stray sideways into parallel-connected loads and their RCDs.

 

[oscar21]

So after reading that other thread would I be correct in saying type A RCD's actually trip if it detects DC current thus disconnecting upstream RCD's from the DC fault anyway? I was under the impression that type A could only cope with DC leakeage as opposed to detect it.

 

[pc1966]

So after reading that other thread would I be correct in saying type A RCD's actually trip if it detects DC current thus disconnecting upstream RCD's from the DC fault anyway? I was under the impression that type A could only cope with DC leakeage as opposed to detect it.

No, type A still detects AC, just that the waveform can have a proportion of DC and it still works. A bigger proportion than that of type AC.

The fault-case to cover is rectified mains, the sort of thing you might get if a switch-mode power supply shorts to earth, or you have a hairdryer that uses half-wave rectification to reduce heat/motor speed, etc. The SMPSU is actually more complicated as for a small load it is near pure DC, but if you draw much current then the ripple is significant.

Where a lot of true DC is available is things like battery storage (e.g. EV) or a PV system if you can get leakage from the panels to the AC side (not always possible, for example an ABB inverter I once dealt with stated type AC is adequate).