'Leak' detection for RCD trip troubleshooting

[bernardgreen]

A couple of points.

[1] We seem to be talking only about appliances tripping the RCD. It may appear that it is when an appliance is operating that the the RCD trips. That does not necessarily mean it is the appliance at fault. The load current of the perfectly good appliance together with an insulation fault elsewhere may be the reason the RCD trips. The "insulation fault" may be the capacitors between neutral and CPC in the mains filters of some other appliance.

[2] Original RCDs were mechanical and had no electronics. The output currrent from the senser toroid was directly proportional to the difference in Live and Neutral through the sensor. This current was used to energise a small electro-magnet that un-latched a delicate spring loaded precision mechanism that forced the contacts apart. Each unit had to be adjusted during manufacture to ensure tripping occured at the specified difference between Live and Neutral, this made them expensive to manufacture. Many modern ones use electronics to avoid having to have precision mechanical mechanisms and to enable one design to be used for diffferent types of RCD. Different trip currents and different delay before operation.

[3] Electronics require power to function. It does seem that some RCDs may not operate if there is no Neutral to the RCD even though there is leakage from Live to earth in the protected circuit. This may be the reason that RCBOs have an "operational earth" which is probably to ensure the electronics are powered when there is Live but no Neutral on the input.

[4] Somewhere I saw a schematic of an RCD with two sensor toroids. One feeding an electronic module and the second feeding direct to an electro-magnet trip. Possibly the second is a back up in case the electronic trip is unable to function.

Personally I would prefer to pay more for a non electronic RCD from a reputable source than risk a low cost item dependant on both Neutral and cheap electronics and wide tolerance mechanics.

 

[ericmark]

I did when first reading 314.1 wonder how any designer can work out what is required to comply?

Every installation shall be divided into circuits, as necessary, to:
(iv) reduce the possibility of unwanted tripping of RCDs due to excessive protective conductor currents produced by equipment in normal operation

531.2.4 An RCD shall be so selected and the electrical circuits so subdivided that any protective conductor current which may be expected to occur during normal operation of the connected load(s) will be unlikely to cause unnecessary tripping of the device.

Where items are moveable on 13A sockets then 543.7.1.1 limits the leakage to 3.5 mA plus 612.5.2 adds 1 mA for cable inductance and capacitance.

I can't find where it says half the rated current must not trip the RCD but with the exception of specials like the X-Pole we need to ensure no more than 15 mA will flow so maximum of 4 sockets.

OK 4 sockets per RCD is going a little daft, but so is 100, and there has to be a limit to the number of sockets protected by one RCD. For RCD's in series (531.2.9) there is a NOTE: In such cases the downstream RCD may need to disconnect all live conductors. However 537.1.2 states:

Except as required by Regulation 537.1.4, in a TN-S or TN-C-S system the neutral conductor need not be isolated or switched where it can be regarded as being reliably connected to Earth by a suitably low impedance. For supplies which are provided in accordance with the Electricity Safety, Quality and Continuity Regulations 2002, the supply neutral conductor (PEN or N) is considered to be connected to Earth by a suitably low impedance.

So we have two considerations one is how many sockets to one RCD and the other should it be double pole to prevent one circuit effecting another?

We have been guided more by manufacturers than by regulatory bodies and the old 16th edition still had a reference to dividing into circuits but it did not clearly define it was also with leakage protection as well as over current. Even if we wanted to fit double pole RCBO's to every circuit with the available consumer units this would be hard. OK there are distribution boards but they are not type tested.

If we followed the requirements of 314.1 then the whole problem of working out what had tripped the RCD would not be an issue. My mother has a bank of RCBO's supplying the kitchen and they have never tripped other than when using the RCD tester. My house with just 2 RCD's about once a month on average one will trip.

I was in favour of swapping to an X-Pole RCD for my house but after seeing how my mothers house has never had one trip I am now leaning towards new consumer unit and RCBO's.

Anyway this brings me back to how many sockets can be feed from one RCD and be considered to comply with 314.1? I would consider 20 sockets is about right, but that's 20 sockets not 20 sockets plus the lights and plus the cooker etc.

My house lights no a problem as central light above stair case is battery backed but this is not repeated with other houses in the estate. With a central stair case I would consider it should be lit from at least 2 RCD supplies but my house was never wired with a split lighting circuit one 6A fuse originally fed all lights.

This to me is one of the problems, we take a house designed to use a fuse box and try to change to MCB's and RCD's then wonder why we have problems. I never had a bulb blowing rupturing a fuse but when we went to MCB it would often trip the MCB. So as design (fuse) it was OK to have one supply to all lights. It was my fiddling fitting MCB and RCD protection which caused the problem.

 

[JohnW2]

[1] We seem to be talking only about appliances tripping the RCD. It may appear that it is when an appliance is operating that the the RCD trips. That does not necessarily mean it is the appliance at fault. The load current of the perfectly good appliance together with an insulation fault elsewhere may be the reason the RCD trips. The "insulation fault" may be the capacitors between neutral and CPC in the mains filters of some other appliance.

Exactly – which is why, despite what some have suggested, I don’t think that ‘systematically disconnecting’ appliances is a foolproof way of identifying the cause of an RCD trip. That’s why I’m suggesting that to actually monitor the L-N balance in the feed to the appliance is a better way to go.

[2] Original RCDs were mechanical and had no electronics. The output currrent from the senser toroid was directly proportional to the difference in Live and Neutral through the sensor. This current was used to energise a small electro-magnet that un-latched a delicate spring loaded precision mechanism that forced the contacts apart. Each unit had to be adjusted during manufacture to ensure tripping occured at the specified difference between Live and Neutral, this made them expensive to manufacture. Many modern ones use electronics to avoid having to have precision mechanical mechanisms and to enable one design to be used for diffferent types of RCD. Different trip currents and different delay before operation.

We discussed this recently. Whilst you may be right in your assumption that the inclusion of electronics was “to avoid having to have precision mechanical mechanisms and to enable one design to be used for diffferent types of RCD”, it is also possible that it was done in order to produce greater long-term reliabity of trip threshold. To rely on the mechanical properties and behaviour of an individually adjusted mechanical mechanism remaining unchanged during a service life of possibly many years is asking a lot – whereas the behaviour of an electronic mechanism ought to remain largely unchanged ‘indefinitely’.

[3] Electronics require power to function. It does seem that some RCDs may not operate if there is no Neutral to the RCD even though there is leakage from Live to earth in the protected circuit. This may be the reason that RCBOs have an "operational earth" which is probably to ensure the electronics are powered when there is Live but no Neutral on the input.

We also discussed this recently. It is, indeed, the stated reason for RCBOs having a ‘functional earth’ connection, but the manufacturers of electronically-activated RCDs do not appear to have felt the need to do the same, and there is debate out there as to whether the functional earths are ‘really necessary’. Of course, the functional earth will be of limited use (no use at all if there are no ‘incidental’ paths to earth via bonding etc.) with TN-C-S, since neutral and earth are then the same thing! You are probably worrying about the ‘vanishingly improbable’. It would take only the slightest of connections to supply neutral (and only a very small L-N pd) for the RCD electronics and trip solenoid to work. For there to be a total loss of the neutral connection would be rare enough, so for their to be a simultaneous L-E (or N-E) fault would probably have a probability in the ‘being struck by a meteorite’ ballpark!

[4] Somewhere I saw a schematic of an RCD with two sensor toroids. One feeding an electronic module and the second feeding direct to an electro-magnet trip. Possibly the second is a back up in case the electronic trip is unable to function.

... or, as I suggested above, maybe the other way around. The behaviour of the electronic one could well be ‘more reliable’ in long-term service than one dependent totally on an electromechanical mechanism,no matter how 'high precision' it was when initially manufactured.

Personally I would prefer to pay more for a non electronic RCD from a reputable source than risk a low cost item dependant on both Neutral and cheap electronics and wide tolerance mechanics.

As above, I think I might be inclined in the other direction – but, in the absence of actual data, it has to be a personal judgement (aka guess!).

Kind Regards, John

 

[JohnW2]

I did when first reading 314.1 wonder how any designer can work out what is required to comply? .... Where items are moveable on 13A sockets then 543.7.1.1 limits the leakage to 3.5 mA plus 612.5.2 adds 1 mA for cable inductance and capacitance. ... I can't find where it says half the rated current must not trip the RCD but ... we need to ensure no more than 15 mA will flow so maximum of 4 sockets. OK 4 sockets per RCD is going a little daft, but so is 100, and there has to be a limit to the number of sockets protected by one RCD.

It is essentially impossible to properly design a sockets circuit, since the designer can but guess what will be plugged into them’. It’s not the number of sockets that matters, but the number of loads plugged into them (and switched on) at any one time. I would imagine that, even if there were 100 sockets, in most domestic situations it would be rare for their be more than a handful of ‘active’ loads connected at any point in time.

However 537.1.2 states:

Except as required by Regulation 537.1.4, in a TN-S or TN-C-S system the neutral conductor need not be isolated or switched where it can be regarded as being reliably connected to Earth by a suitably low impedance. For supplies which are provided in accordance with the Electricity Safety, Quality and Continuity Regulations 2002, the supply neutral conductor (PEN or N) is considered to be connected to Earth by a suitably low impedance.

So we have two considerations one is how many sockets to one RCD and the other should it be double pole to prevent one circuit effecting another?

I don’t really understand that paragraph of 537.1.2. It seems to be saying that isolation/switching is not required for any ESQCR-complaint supply. Are there any non-compliant supplies and, if so, how are we meant to identify them?

If we followed the requirements of 314.1 then the whole problem of working out what had tripped the RCD would not be an issue.

Indeed. It goes without saying that the less loads one has protected by each device, the easier will it be to identify the cause of a trip.

My house with just 2 RCD's about once a month on average one will trip.

Maybe I’m just lucky, since I’m a little surprised by what I hear of the experiences of others. I have a bout 10 RCDs (plus a few RCBOs) in my house (many CUs), and I don’t recall a ‘nuisance trip’ of any of them, over a period of a number of years. The very few trips that have occurred have been ‘for good reason’.

Kind Regards, John

 

[EFLImpudence]

It goes without saying that the less loads one has protected by each device...

Tut, tut.

Now if I had recently written "It would be fewer than a large electric shower" I would have expected a lashing.

 

[JohnW2]

It goes without saying that the less loads one has protected by each device...

Tut, tut. Now if I had recently written "It would be fewer than a large electric shower" I would have expected a lashing.

Indeed - guilty as charged

Kind Regards, John

 

[Aragorn84]

Personally i would view it somewhat unsatisfactory to have any genuine "nuisance" trips, ie ones without being caused by an actual fault.

This current place we've been in for 4 years is the first property i've lived in which has an RCD board, and the only times the RCD has tripped that i can recall, was due to some shonky underfloor junction boxes getting wet due to a high water table under the house and lots of rain. A nuisance perhaps, insofar as at the time it was annoying and had no obvious cause, but a genuine fault/trip nonetheless, and once discovered and rectified, there has been no further issues.

 

[JohnW2]

Personally i would view it somewhat unsatisfactory to have any genuine "nuisance" trips, ie ones without being caused by an actual fault.

Me, too - but from what one hears (in this formum and more widely), it's a pretty common occurrence.

This current place we've been in for 4 years is the first property i've lived in which has an RCD board, and the only times the RCD has tripped that i can recall, was due to some shonky underfloor junction boxes getting wet due to a high water table under the house and lots of rain. A nuisance perhaps, insofar as at the time it was annoying and had no obvious cause, but a genuine fault/trip nonetheless, and once discovered and rectified, there has been no further issues.

As I said, same here. I've had 10+ RCDs for many years, and the only (IIRC, about 3) trips I can recall were due to 'genuine faults', at least one being similar to the one you describe ... so I don't know if I'm just lucky, or what.

Kind Regards, John.