6 mA DC testing and 9 mA AC testing with reference to RCD's

[ericmark]

Pure interest, as I don't have solar panels or an EV charging point, but I read on another forum "all info refers to using a 1741+ that has a specific 'EV' setting for going the 6mA DC test." sure it should say doing.

I assume it injects 6 mA DC to see if it auto disconnects? As why only EV charging points have a 6 mA DC or a loss of PEN detection I don't know, as any diode can cause DC, and any item outside of the equipotential zone can cause danger with loss of PEN, it is not restricted to EV charging points.

I tried to measure DC with my own house, and found the zeroing of the clamp on means to measure 6 mA is not easy, with the AC scale measuring the leakage in mA to ensure under 9 mA not a problem, but DC seems a little hit and miss.

So putting a 6 mA DC load does seem a reasonable method, I look at this and the three port valve has been around for some time, no idea how much goes through the motor, but these must cause some DC on the supply, but not seen any directives to the effect any property with Y plan should fail an EICR with type AC RCD's.

I personally take the attitude the RCD is secondary protection, and with 14 RCBO's it is very unlikely a fault will stop the RCD bit working at the same time as a earth fault happens. The RCBO's all said type B on the boxes, it was only after they were fitted found they were type AC.

I found this thread started by @RF Lighting back in 2016 talking about testing for mA leakage, and back in 2004 I know I had a problem with tripping RCD which seemed to be cured by replacing it, although the RCD tester showed both old and new to be well within spec.

Back then never even considered DC. In 2010 @bernardgreen did some tests here with interesting results.

So it seems nothing new, but we still test the RCD with nothing connected, and until last year I could not measure DC mA or and AC mA under 10 mA, and I am sure many are the same, the insulation tester uses DC so that not really any help.

When I started using a loop impedance tester on a regular basis I found I got a feel for likely results, so although it may have passed, I would question it when unexpected results were found, but today I don't do enough testing to get a feel with AC and DC milliamps with regard to RCD's.

 

[JohnW2]

Pure interest, as I don't have solar panels or an EV charging point, but I read on another forum "all info refers to using a 1741+ that has a specific 'EV' setting for going the 6mA DC test." sure it should say doing. .... I assume it injects 6 mA DC to see if it auto disconnects? As why only EV charging points have a 6 mA DC or a loss of PEN detection I don't know, as any diode can cause DC, and any item outside of the equipotential zone can cause danger with loss of PEN, it is not restricted to EV charging points.

Although somewhat tangential to what you are saying, I still have only a very limited understand of the behaviour (required behaviour) of the various 'Types' of RCDs we so often talk about. Of 'Type A' ones, BS7671 says ...

(ii) RCD Type A: RCD tripping on alternating sinusoidal residual current and on residual pulsating direct current,
suddenly applied or smoothly increasing.

In the absence of any comment to the contrary, I assuming that 'tripping referred tio is that which occurs if the residual current exceeds the IΔn of the device - so typically 30mA. If that is the case, then it seems to be saying that tripping "on residual pulsating direct current, suddenly applied or smoothly increasing" if that residual current is ≥30mA - do you agree? However, it then goes on to add this note...

NOTE 1: For RCD Type A, tripping is achieved for residual pulsating direct currents superimposed on a smooth direct current up to 6 mA.

... which, if I'm correct in my interpretation above, presumably still means that it will trip with "residual pulsating direct currents" ≥30mA, even if there is a superimposed 6mA 'smooth direct current'.

Do you, or others, think that is a correct interpretation?

Kind Regards, John

 

[ericmark]

I looked at the RCD's I fitted in 1992 OK second hand when I fitted them, but there is nothing about type AC, A, F or B and I don't have a clue which they are.

I consider they are secondary protection, and with a TN supply, not really worried, if as did happen I get a water leak, which goes into a socket, the RCBO will likely trip before I touch anything live. Even if it takes ½ hour to trip, as long as it trips before I touch it, then its done its job.

However there is no cut out or switch on my earth cables, so loss of PEN has nothing which can protect me with anything used outside the equipotential zone, except most things I use outside are class II.

I know after hitting a cable out of the permitted zones, even with 30 mA RCD's the shock can knock one out, 40 mS is too long to get the shock for, what we hope is not that it trips when some one touches a live cable, but that it will trip before some one touches it, as when sockets get wet.

But 6 mA is a very small DC current, and we have so many items today which could cause 6 mA or more to flow, without some auto disconnect device, it in real terms does not matter if type AC or type A, there a good possibility it will not trip within the 40 mS. If the type A could take 600 mA DC then yes worth the change from AC to A, but at 6 mA there is so little difference it makes no never mind.

I don't know if you have tried to measure it, I did, this will show 6 mA on a cable away from magnetic forces when first zeroed, but clamped around the main tail out of the DNO fuse, it is so hard to zero, it is a bit hit and miss measuring 6 mA.

I assume the tester actually inserts a 6 mA DC component to see if the detection device works? But with no solar panels or EV charging points fitted I have no 6 mA DC detection device fitted, and nothing to auto open the supply if outside the 207 to 253 volt range, and I would not really want one, as that would mean with the shortest power cut my freezers would switch off and stay off until I manually reset the device.

And except for Algeria with generator supplies three phase to accommodation modules, I have only come across lost of PEN once. So not really worried about it.

 

[JohnW2]

I looked at the RCD's I fitted in 1992 OK second hand when I fitted them, but there is nothing about type AC, A, F or B and I don't have a clue which they are. I consider they are secondary protection, and with a TN supply, not really worried, if as did happen I get a water leak, which goes into a socket, the RCBO will likely trip before I touch anything live. Even if it takes ½ hour to trip, as long as it trips before I touch it, then its done its job.

Unless the earth leakage 'gets worse' (i.e. earth fault current increases) it will operate (in response to residual current) eithe immediately or 'never' - it's not like thermal tripping of MCBs when an excessive current less than the magnetic tripping threhold may take half and hour, or longer, to trip the device 'thermally'.

However there is no cut out or switch on my earth cables, so loss of PEN has nothing which can protect me with anything used outside the equipotential zone, except most things I use outside are class II.

That's surely the important point, in practice, isn't it? I'm not sure when I last saw anything likely to be used 'outside' which was not Class II, are you? The greatest (theoretical) risk probably relates to the infamous 'outside taps', but the probability of someone touching such a tap (when having bare feet) at the same time as there was a 'lost PEN' (itself very rare) must be 'vanishingly small' mustn't it?

I know after hitting a cable out of the permitted zones, even with 30 mA RCD's the shock can knock one out, 40 mS is too long to get the shock for, what we hope is not that it trips when some one touches a live cable, but that it will trip before some one touches it, as when sockets get wet.

Quite so. Apart from anything else, how much a threat to life is presented by an electric shock is determined by both the magnitude and duration of the current flow through the victim's body, and a residual current device can do nothing about magnitude

But 6 mA is a very small DC current, and we have so many items today which could cause 6 mA or more to flow,

Indeed. I don't really know where this 6mA figure came from (i.e,. what is the thinking behind it).

As I wrote, in relation to standard Type A 30 mA RCDs/RCBOs, my (seemingly 'literal') interpretation of what BS7671 says is that the device is not required to trip with residual currents less than 30 mA, even if they are DC currents - but no-one has yet indicated whether they agree with my interpretation.

.... I have only come across lost of PEN once. So not really worried about it.

Indeed - as I said above, very rare. I would imagine that most people could liver for sn entire normal!) lifetime without every have encountered such an event.

Kind Regards, John

 

[bernardgreen]

the infamous 'outside taps', but the probability of someone touching such a tap (when having bare feet) at the same time as there was a 'lost PEN' (itself very rare) must be 'vanishingly small' mustn't it?

Small enough that a gas meter reader wore gloves after having received a serious electrical shock from an external gas meter.

I have only come across lost of PEN once. So not really worried about it.

In the thirty years at the last house we had at least 3 incidents where the potential between the PME from the network and the ground rod was high enough to be of concern.

One cause was the village sewage pumping station that was supplied from the domestic network going faulty and creating a very large un-balance on the three phases. ( hat accelerated the provision of a dedicated supply from the 11kV network ).

Another cause was a snapped Neutral in the overhead section of the network

 

[bernardgreen]

Another consideration is when a damp and thus slightly conductive wall "imports" ground potential into the house.

 

[ebee]

a 30mA RCD if within spec might be expected to save about 95% of the population but that still leaves 5% it does not save, couple that with perhaps as high as 7% RCD failure rate and it becomes a little more worthwhile thinking about. Plus, even a working RCD can allow quite a shock until it disconnects, you might not get electrocuted (Electrical Execution) but you might fall off ladders/stairs example and break yer neck.

 

[ericmark]

I have seen devices to monitor the bonding to true earth and testers for them which reading the instructions looked for 40 volts to trip the ELCB-v, however this have been discontinued, and I see a real danger with anything which disconnects the earth. Even if it does disconnect the lives first.

Bring back TN-S and TT, the TN-C-S has a problem and there is no real answer other than stopping the use of TN-C-S specially with EV charging.

But having a detector to alert if there is a loss of PEN is easy enough, but to auto disconnect is problematic as it could disconnect the earth in error. One end to earth terminal

and other to earth rod would show when there is a problem.

As to 6 mA DC LED's, mid position motorised valves, inverters or switch mode power supplies, old battery chargers, can all produce DC, it does not need to be an EV or solar panels, and when I used my clamp on it showed around 14 mA, but hard to zero, so not sure if correct, but the test rigs to show the problem with DC

uses 50 mA and 250 mA so 6 mA is nothing really.

 

[flameport]

Type AC are designed for sinusoidal waveforms only. Connect them to anything else and their behaviour is undefined.
Type AC includes those which are marked as such, and older devices such as those made to BS4293.
As the vast majority of electronic equipment does not draw a sinusoidal current, Type AC are next to useless. That being the reason that many countries banned the use of them years or decades ago.

Type A will operate with non-sinusoidal waveforms, which includes 'pulsed DC' or in other words anything from a rectifier or even a single diode.
They do not detect smooth DC but will still operate if up to 6mA of smooth DC is present.
If more than 6mA of smooth DC is present, their behaviour is undefined.

For both AC and A 30mA RCDs, the current imbalance must be between 15mA and 30mA to trip.

The video demonstrates what happens when smooth DC is applied. The RCD used is a Type AC, but the same problem will occur with Type A.

If you have a circuit which only uses sinusoidal AC and there is no possibility of anything else, Type AC is suitable.
For circuits which will have non-sinusoidal waveforms, Type A is the minimum required. The vast majority of modern equipment is in this category.
For any situation where smooth DC is likely, the minimum is Type B.

 

[JohnW2]

Small enough that a gas meter reader wore gloves after having received a serious electrical shock from an external gas meter.

You often tell that story, but I didn't realise it happened at the very moment when there was a 'lost PEN'. Was that really the case?

In the thirty years at the last house we had at least 3 incidents where the potential between the PME from the network and the ground rod was high enough to be of concern.

Only the last example you cite (snapped overhead neutral) is an example of a 'lost PEN' - but, even if you have been unlucky enough to experience that very rare event (which most people will never experience), I still reckon that the probability of someone with bare feet touching an outside tap/whatever at the very time such a fault is present is 'vanishingly small' - even if your gas meter reader is 'an exception which proves that rule'!

You talk of 'PDs of concern' but they are, of course, not of any concern (not a risk to) people within a properly wired electrical installation - any risk to persons only exists outside of the building whilst such a fault is present.

Kind Regards, John

 

[JohnW2]

Type A will operate with non-sinusoidal waveforms, which includes 'pulsed DC' or in other words anything from a rectifier or even a single diode.
They do not detect smooth DC but will still operate if up to 6mA of smooth DC is present.
If more than 6mA of smooth DC is present, their behaviour is undefined.

Thanks for confirming my interpretation of what BS7671 says.

Kind Regards, John

 

[JohnW2]

a 30mA RCD if within spec might be expected to save about 95% of the population but that still leaves 5% it does not save, couple that with perhaps as high as 7% RCD failure rate and it becomes a little more worthwhile thinking about.

As is often said, the primarily value of RCDs in 'saving lives' relates to their ability to clear faults before anyone has a chance to suffer a shock, even when the fault current is far too small for the fault to be cleared by traditional ADS.

Once someone gets a shock, we are into the territory of uncertainties. Over the past good few years I have taken many opportunities to ask groups, sometimes large groups of people (either 'in person' or 'electronically) if anyone has suffered and survived an electric shock which caused an RCD to trip, and have found virtually no examples of that. There undoubtedly atre some examples, somewhere, but even that affords no certainly that they would not have survived in the absence of an RCD. Conversely, if someone has died as a result of a shock from a non-RCD-protected circuit, there is clearly no way of knowing whether or not they would have survived if there had been RCD protection

As you imply, when a shock has happened, there is no certainty about how much 'protection' will be afforded by an RCD to a particular person. For a start, and RCD can do nothing to reduce the magnitude of current through a person, so if that current is high enough, it may well result in a lethal disturbance of heart rhythm before an RCD operates (limiting the duration of the shock, but 'too late'). Furthermore, there are a good few people around (many of whom are unware of this problem) whose hearts are so 'electrically unstable' that just a very small number of mA for a few milliseconds (too little to trip an RCD) may be enough of a 'last straw' to result in a lethal rhythm disorder.

Kind Regards, John

 

[bernardgreen]

You often tell that story, but I didn't realise it happened at the very moment when there was a 'lost PEN'. Was that really the case?

All I know is that which the meter reader told me, He had received an electric shock when he touched the gas meter, Obviously the CPC in the house and ( one assumes ) bonded to the gas pipework ) was at a different potential to ground.

very rare event

Maybe not so rare. after all it will go unnoticed if no one gets a shock or damage is caused.

Snapped Neutrals.
Neutrals disconnected by metal theft,
high currents in the network Neutral due to phase load unbalance.

 

[JohnW2]

All I know is that which the meter reader told me, He had received an electric shock when he touched the gas meter, Obviously the CPC in the house and ( one assumes ) bonded to the gas pipework ) was at a different potential to ground.

Yes, presumably, but I very much doubt that was due to a PEN fault.

Maybe not so rare. after all it will go unnoticed if no one gets a shock or damage is caused.

I doubt that a true ('complete') 'lost PEN' will often go unnoticed, since it's probably unlikley that the 'incidental paths to earth' (i.e. via bonded extraneous-c-ps) in the properties downstream of the fault will often be of sufficiently low impedance that all equipment (ncluding lights etc.) in the properties will 'function normally' (if at all?

Kind Regards, John

 

[flameport]

It's very likely that open PEN conductors go unnoticed.
The path via common bonded ECPs will inevitably be of very low impedance, as it will be bonded at every installation.

Consider a typical street with 15 properties, all with a common metal gas service bonded in all of the properties.
PEN conductor fails at a point which results in 10 of those properties not having any neutral connection.
All of the neutral current for those 10 properties will now return via the individual bonding conductors, through the metal gas service and through the bonding conductors of the 5 properties which are still connected to the PEN.
Everything works as normal, no one suspects anything until bonding conductors in the remaining properties get hot, gas meters start to melt and so on. With low or moderate loads it could be weeks or months before anything seems wrong.