RCBO keeps tripping

[Peet]

Hmmnn. I'm sure that at least once I've seen an RCD trip for a N-E short with the RCD unloaded. I don't get involved with various types of electrical installations at the rate you chaps do, and rarely have I been involved with a TT system. So I'd say I've seen it with a TNS system, and not PME. It stuck in my mind because all the MCBs where off at the time. But at the same time it didn't seem strange to me due to the low impedance circuit and long cable runs present, although the neutral would have been overall shielded by the PE. If it had of been PME then it really would of stuck in my mind due to the local link between the N & E.

The other thought about the OP's problem (and possibly my previous one) is - do RCDs develop faults whereby they loose their ability to sense differentially? In electronic terms, depending on its design the circuitry (if indeed they are electronic) may have to reject the effects of large static or transient currents to be able to sense the small differential in the line and neutral. In the electronics world this would be called 'common mode rejection'. A failure or imbalance in the device would cause larger loads (with identical line & neutral currents) to be seen as a differential imbalance, so tripping the device. And would give rise to what the OP sees.

Depending on the circuitry problem the false trip could be caused by a either a large static current or a large transient.

To help the OP, my thoughts would be to check the neutrals are correctly placed and then get somebody in to test the RCD and/or swap it out.

 

[EFLImpudence]

do RCDs develop faults whereby they loose their ability to sense differentially?

They may but then they are just 'broken'.

They should be tested quarterly but people don't and then one day they don't work.

In electronic terms, depending on its design the circuitry (if indeed they are electronic) may have to reject the effects of large static or transient currents to be able to sense the small differential in the line and neutral.

RCBOs - I wouldn't think so. RCCBs - definitely not.

 

[JohnW2]

Hmmnn. I'm sure that at least once I've seen an RCD trip for a N-E short with the RCD unloaded. I don't get involved with various types of electrical installations at the rate you chaps do, and rarely have I been involved with a TT system. So I'd say I've seen it with a TNS system, and not PME.

As Bernard said, it's certainly theoretically possible- so you may have seen it - although, as I said (and as you say) probably not with TN-C-S/PME. Like Bernard, I'm not an electrician, so have not had that much opportunity to see it. The point obviously is that if there is no load, there should not be current flowing through either side (L or N) of the RCD. For an imbalance to occur as a result of an N-E fault, that means that some current has to flow through the N side of the RCD (the L side presumed to be zero) - and, in the absence of load, the only conceivable way that could happen would be if it was flowing from the N to 'somewhere' - as Bernard said, presumably 'true earth'.

The other thought about the OP's problem (and possibly my previous one) is - do RCDs develop faults whereby they loose their ability to sense differentially? In electronic terms, depending on its design the circuitry (if indeed they are electronic) may have to reject the effects of large static or transient currents to be able to sense the small differential in the line and neutral. In the electronics world this would be called 'common mode rejection'.

When they are working 'as intended, RCDs obviously have very high 'common mode rejection' - since, say, an 8 A/30 mA one has to be able to recognise 80.000 A vs 80.000 A as balanced, but 80.030 A vs 80.000 A as imbalanced. If/when they develop faults, it's obviously possible for anything to happen. However, since the differential sensing is done electromagnetically (just with a toroid), I doubt whether a change in that, per se, is a likely failure mode.

Kind Regards, John

 

[Peet]

This is coming away from the original problem (Sorry OP!) but its got me thinking.

Folks, check my thinking on this one would you, as to how an unloaded RCD could trip. If the unloaded RCD is still in the board, the supply is TNS and other non RCD circuits are producing a neutral current then when the neutral on the output side of the RCD was shorted to earth the original neutral current would then split between the earth and the neutral. The earth current would be through the RCD and since it's live side has zero current a trip would come at some point just above the nominal 30mA depending on the ratio of currents, if the device is purely electromagnetic in its operation. I'd guess that PME could do it too but at a much elevated neutral current.

 

[EFLImpudence]

I'd guess that PME could do it too but at a much elevated neutral current.

No, because the N & E are connected, i.e. negligible impedance.

On TNS, two houses are no different from two RCD sides in one board as the N & E are not connected until back at the transformer.
That's why I said the impedance is relatively high, i.e. not negligible.

Because of this, is (small) voltage not on the neutral all the time which could cause the problem as in Bernard's 'theory'?
Indeed, is it a problem?

However, the problem is not likely to occur as an RCD is unlikely to have NO load on it and so it would have already tripped.

Were it to happen then an electrician should, with the necessary testing, discover the cause and so the RCD would have discovered and disconnected a fault - as it should.

 

[JohnW2]

Folks, check my thinking on this one would you, as to how an unloaded RCD could trip. If the unloaded RCD is still in the board, the supply is TNS and other non RCD circuits are producing a neutral current then when the neutral on the output side of the RCD was shorted to earth the original neutral current would then split between the earth and the neutral. The earth current would be through the RCD and since it's live side has zero current a trip would come at some point just above the nominal 30mA depending on the ratio of currents ....

Yes, I think that's correct. When I talked about an 'unloaded RCD', I was actually thinking in terms of one RCD covering the whole installation - I should have said 'unloaded installation'. As you say, in non-TN-C-S systems, some of the neutral current from non-RCD-protected circuits may flow ('backwards') through through the N side of an RCD if there is a N-E fault on the load side of the RCD, thereby causing the ('unloaded') RCD to trip.

Kind Regards, John

 

[Electrifying]

Yes, I think that's correct. When I talked about an 'unloaded RCD', I was actually thinking in terms of one RCD covering the whole installation - I should have said 'unloaded installation'. As you say, in non-TN-C-S systems, some of the neutral current from non-RCD-protected circuits may flow ('backwards') through through the N side of an RCD if there is a N-E fault on the load side of the RCD, thereby causing the ('unloaded') RCD to trip.

Kind Regards, John

In theory, is that not just as likely to happen even with a 'non-loaded installation' - regardless of 'earthing system'.

My thinking:

All MCBs off so no load.
Neutral-earth fault on one circuit.
A neighbour's installation pulling 40A load.......some of that must travel up your neutral, through the RCD and to earth via the fault.

It's actually probably less likely to happen on a TN-C-S.

 

[JohnW2]

In theory, is that not just as likely to happen even with a 'non-loaded installation' - regardless of 'earthing system'. My thinking:
All MCBs off so no load. Neutral-earth fault on one circuit. A neighbour's installation pulling 40A load.......some of that must travel up your neutral, through the RCD and to earth via the fault.

Try as I may, I can't fault the logic of that! However, I have to say that my limited anecdotal experience is that N-E faults (accidentally touching N and CPC together!) generally do not seem to trip RCDs if all the MCBs are off. However, given that I can't fault your logic, that's hard to explain - so maybe there's something wrong with my recollections/experience.

It's actually probably less likely to happen on a TN-C-S.

That's what I said - all of what I've written recently has been prefaced/qualified by "...in non TN-C-S systems". As far as I can see, with TN-C-S and a proper equipotential zone (i.e.adequate bonding), what you describe should not really happen with an N-'E' fault (i.e. N-MET/CPC) within the equipotential zone, since N and 'E' are joined at the origin of the installation, so there should be no PD to push any current through the N side of the RCD when, as in the case you describe, the supply side of the RCD is connected to N and the load side (via a fault) to 'E'. As far as I can see, only if (as per what Bernard was saying, I think) there is a fault between N and 'true earth' (and if that is at a different potential from N/MET/CPCs) could it happen with TN-C-S - and that situation should not arise within the premises with adequate bonding. It might, of course, happen with the infamous 'garden tools'!

Kind Regards, John.

 

[bernardgreen]

and that situation should not arise within the premises with adequate bonding. It might, of course, happen with the infamous 'garden tools'!.

Even with all the bonding in place a neutral to CPC fault could still trip an RCD if the CPC is in any way connected to true ground in the property..

If the in-coming neutral is not at ground potential then the CPC is not at ground potential and current will flow in the CPC to ground. The neutral to CPC fault provides a parallel path from the incoming neutral to the CPC and this parallel path is through the sense coil of the RCD.

 

[JohnW2]

and that situation should not arise within the premises with adequate bonding. It might, of course, happen with the infamous 'garden tools'!.

Even with all the bonding in place a neutral to CPC fault could still trip an RCD if the CPC is in any way connected to true ground in the property.. If the in-coming neutral is not at ground potential then the CPC is not at ground potential and current will flow in the CPC to ground. The neutral to CPC fault provides a parallel path from the incoming neutral to the CPC and this parallel path is through the sense coil of the RCD.

I don't really understand what you are saying. I was specifically talking about TN-C-S and a totally unloaded installation. With TN-C-S, incoming N is connected to the MET (hence the installation's CPCs) at the origin of the installation. Connecting the CPCs (hence incoming N) to some additional 'true earth' would still leave one with an equipotential zone. Indeed, that additional connection to 'true earth' would then become nothing more than a further part of the PME earthing. I don't see why changing the 'M' of PME from, say, 6 to 7 would make any difference to anything.

I suspect you must be talking about something other than what I have (mis?) understood from you post.

Kind Regards, John

 

[roasty]

As a kid, I had both a Commodore Amiga (A500) and a plug in rcd adaptor for use when I was doing any electrical 'experimenting' - my parents got it for me to increase safety.......

Anyway, I soon realised that regardless of where in the house this rcd adaptor was plugged, as long as it was reset (I.e ready for use) you could trip in by simply plugging in the Amiga power block anywhere else in the house and flicking the power switch!

To this day, I'm not sure why. But it taught me that there isn't always a reason why rcds trip.
Dan

 

[JohnW2]

As a kid, I had both a Commodore Amiga (A500) and a plug in rcd adaptor for use when I was doing any electrical 'experimenting' - my parents got it for me to increase safety.......

When I was a kid, my parents just had to worry - RCDs hadn't been invented!

To this day, I'm not sure why. But it taught me that there isn't always a reason why rcds trip.

Oh - there's always a reason, that's for sure. It's just that sometimes none of us are clever enough to work out what it is

Kind Regards, John

 

[bernardgreen]

Connecting the CPCs (hence incoming N) to some additional 'true earth' would still leave one with an equipotential zone.

Almost equipotential, a few millivolts of potential difference can still exist between ends of bonds and earth wires. Those few millivolts produce a current that with the milliohm impedances can be surprisingly high

Indeed, that additional connection to 'true earth' would then become nothing more than a further part of the PME earthing.

Yes but between this ground connection and the neutral is the sensor of the RCD

I don't see why changing the 'M' of PME from, say, 6 to 7 would make any difference to anything.

The difference is that the route to the ground is a pair of parallel paths from the MET to ground. One of those routes is via the neutral through the RCD, along the neutral to the point where neutral faults to ground and the CPC neutral to gound .

 

[JohnW2]

Connecting the CPCs (hence incoming N) to some additional 'true earth' would still leave one with an equipotential zone.

Almost equipotential, a few millivolts of potential difference can still exist between ends of bonds and earth wires. Those few millivolts produce a current that with the milliohm impedances can be surprisingly high

Indeed, that additional connection to 'true earth' would then become nothing more than a further part of the PME earthing.

Yes but between this ground connection and the neutral is the sensor of the RCD

I don't see why changing the 'M' of PME from, say, 6 to 7 would make any difference to anything.

The difference is that the route to the ground is a pair of parallel paths from the MET to ground. One of those routes is via the neutral through the RCD, along the neutral to the point where neutral faults to ground and the CPC neutral to gound .

I must be misunderstanding you somewhere, or missing something, because I'm struggling to get may head around that. The above all refer back to your statement "Even with all the bonding in place a neutral to CPC fault could still trip an RCD if the CPC is in any way connected to true ground in the property". "Connecting the CPC [hence MET - and N, with TN-C-S) to true ground in the property" is all on the supply side of any RCDs and, as I said, is seemingly no different from any of the other PME connections from N to (a version of) 'true earth' outside of the property, is it?

Kind Regards, John