Neutral to earth fault affecting both RCDs

[Roadhousezamora]

Hello everyone, I'm newly registered having read the boards from the shadows for a while. I have a question about neutral to earth faults and RCDs. I have read around this as much as I can but being relatively newly qualified my experience isn't as much as the collective members here so I thought I'd ask.

I had been asked to install a new extractor fan in a kitchen, I have done some pre work testing on the cabling present and it all looked good so I proceed to fit the extractor. When filling out the minor works cert I have struggled to get the Rcd to trip. I know I should have tested this as well before I started the work but as I said I'm new and clearly a bit naive. Initially the Rcd auto test didn't trip at 1/2 on both sides of the wave, tripped the Rcd both times at 1x then gave me a 50v error at 5x, the test button worked.

I did a bit of reading and found that the fault could be a neutral to earth fault. I pulled out each neutral at the board and IR tested between individual neutrals and the earth bar of all the circuits on that side of the board and found no problems >200M on each circuit. When carrying out a global IR test (neutral bars to earth bar) I see there is a fault but on the other side of the board. I pulled all the neutrals out on the other side and identified a neutral to earth fault on the sockets ring upstairs. I have chased this to a cable between two sockets and will obviously now have to replace this but my question was, can a neutral to earth fault on the other side of the board affect both RCDs? Was the 50v warning on the rcd I was initially testing due to the neutral to earth fault I have identified on the other side of the board?

From my reading I have seen that this type of fault can cause the RCDs to not trip at all and wondered if the fault could be affecting my ability to test this RCD?

Any help would be appreciated and any additional info will be provided swiftly.

If it helps anyone the RCDs are actually Lewden type A RCCBs (63A 30mA)

Thanks

 

[ImpededLoop]

Hi there and welcome.

What's the ZE and supply type (TT, TNS or TNCS)?

 

[Roadhousezamora]

Hi there and welcome.

What's the ZE and supply type (TT, TNS or TNCS)?

TNS - 0.24

 

[timhoward]

Initially the Rcd auto test didn't trip at 1/2 on both sides of the wave, tripped the Rcd both times at 1x then gave me a 50v error at 5x, the test button worked.

Hi and welcome.
Just checking you know it's not supposed to trip at 1/2.
The 50v error at x5 indicates you have not got a supply voltage at the output terminals and the test cannot proceed; this can be that the 'jump' of the trip slightly shook one of the probes off.
It might be worth trying that one again.

It might be worth pointing out that regulation 415.1.1 says that an RCD must operate within 40 ms when tested at a current equal to or higher than five times its rated residual operating current.
So if you get <40ms at X1 (which invariably you do these days) it's already in fact passed.

I'm just reading and digesting the other bit - well done for finding the N-E fault - they can be a nightmare to find.

 

[Roadhousezamora]

Hi, thanks, it took a while ? I'm more concerned with how to fix it without having to rip the guys wall down, I only went in to put a new extractor up ?
yeah with regards the 1/2 (15mA) test I was getting 2000ms and no trip ?

 

[timhoward]

I did a bit of reading and found that the fault could be a neutral to earth fault. I pulled out each neutral at the board and IR tested between individual neutrals and the earth bar of all the circuits on that side of the board and found no problems >200M on each circuit. When carrying out a global IR test (neutral bars to earth bar) I see there is a fault but on the other side of the board. I pulled all the neutrals out on the other side and identified a neutral to earth fault on the sockets ring upstairs. I have chased this to a cable between two sockets and will obviously now have to replace this but my question was, can a neutral to earth fault on the other side of the board affect both RCDs? Was the 50v warning on the rcd I was initially testing due to the neutral to earth fault I have identified on the other side of the board?

Assuming the RCDs are working properly, in normal circumstances a N-E fault on one side of a dual RCD board should only affect that side.
I'm wondering if as you've had most wires out of the board to fault find (sensibly!) then is there a possibility one of the Neutrals is now in the wrong bar, which is very easily done.

I only went in to put a new extractor up ?

From time to time this sort of thing happens and its frustrating for you and the customer. I'm not trying to be patronising but you are handling it very well, most newly qualified people don't cope too well when testing goes off-script and the wrong number appears, and you've worked out the fault and found it. Don't beat yourself up!

What I'd do next:
I'd be inclined to ramp test both RCDs ideally with all breakers off and the outgoing N disconnected to check they are both tripping at 20-30ma with no loads.
I'd probably remove the N's for faulty socket circuit at the CU, leave that breaker off, and test the RCDs normally again and see what happens.
You could also simulate a N-E fault between each N bar and CPC bar at the CU and see if both RCDs still trip.

 

[Roadhousezamora]

Thanks for the advice, I really appreciate it. All neutrals are in the right places as I removed them one at a time to find the circuit. With regards the ramp testing and testing in general across the Rcd without any loads sounds like a good idea. I have bought a new RCD as both are over 11 years old so I can't rule out a failure of both RCDs due to age but I will return that if the testing without loads shows them to be operating correctly. My instinct is that the Rcd on the side I was working on has failed and the Rcd on the neutral earth fault side is not operating due to a fault, the fact they have occured together is hopefully coincidence but I won't be taking anything for granted with it. I can't get back there till Sunday so there's plenty more time for me to consider other possibilities.

 

[telectrix]

it has been known for a N-E fault on RCD1 circuits to trip RCD1 when a load is switched on on RCD2. this is due to some witchcraft of current flowing down the wrong path. even had 2 RCDs once would not trip either by tests or with the button. was a hard N-E on a redundant but still connected immersion heater circuit.

 

[loz2754]

I find it curious that the OP states that the RCD showed 50V error warning on the 5x test. This indicates that the tester is not seeing the required voltage to perform the test.

 

[timhoward]

I find it curious that the OP states that the RCD showed 50V error warning on the 5x test. This indicates that the tester is not seeing the required voltage to perform the test.

Agreed. If it's an RCD with deep slots on the terminal screws it can be the GS38 tips prevent proper contact especially if you aim in the middle. I suspect a probe slipped slightly.
Obviously checking for loose connections is wise too!

it has been known for a N-E fault on RCD1 circuits to trip RCD1 when a load is switched on on RCD2. this is due to some witchcraft of current flowing down the wrong path. even had 2 RCDs once would not trip either by tests or with the button. was a hard N-E on a redundant but still connected immersion heater circuit.

Interesting. I wonder if it could be explained by a load with a reasonable inrush current on RCD2 causing a momentary volt drop, in turn causing a momentary current draw rise on RCD1 pushing the leakage current over the threshold...
I almost went for a dual rcd board for an install next week where I'm travelling a fair distance and can't get good information about current circuits other than there are lots of them. My thinking had been I can take lots of MCB's with less outlay than lots of RCBOs. In the end I gave myself a good talking to and went the RCBO route, and this kind of thing reassures me it was the right call!

 

[Roadhousezamora]

I ran the test using the plug top on two different socket outlets on the ring main on the same side as the Rcd I was testing as well as from the extractor. Same 50v error ?

I wondered if it may be some capacitance from a knackered alarm battery which is a fused spur off the ring?

 

[timhoward]

I ran the test using the plug top on two different socket outlets on the ring main on the same side as the Rcd I was testing as well as from the extractor. Same 50v error ?

Ooh hang on. We might be barking up wrong tree.
Do you remember if it was ">50v"?
What tester is is?

 

[Roadhousezamora]

Ooh hang on. We might be barking up wrong tree.
Do you remember if it was ">50v"?
What tester is is?

It's a Dilog 9110 (can't afford anything premium right now lol) it's ok but prone to being temperamental.

 

[ELECNEWT]

I ran the test using the plug top on two different socket outlets on the ring main on the same side as the Rcd I was testing as well as from the extractor. Same 50v error ?

Test instruments meeting the requirements of BS EN 61557 must perform a pre-check when testing to ensure that the earth potential does not rise to a dangerous level during the test. This will apply to earth loop impedance testing and RCD testing since these tests involve producing a current on the earth. If the test current is expected to produce a dangerous voltage, the tester should inhibit the test and show a warning, such as >50 V.

 

[timhoward]

It's a Dilog 9110 (can't afford anything premium right now lol) it's ok but prone to being temperamental.

Nowt wrong with that.
This is where the devil is in the detail:


(On some other makes that message can mean something else.)
For the little time it takes, I'd see if it's now fine with the dodgy socket circuit's neutrals removed.
If still a problem it's IR testing each circuit again I'm afraid, start with L+N together to CPC.

 

[Aaron b]

Are you testing the RCD at the board or on a circuit? No experience with meggers (and we don't install RCD's often for utilities) but I would have assumed that the test had failed because the voltage on the earth conductor was at 50V. I assume the cable resistance was too high to achieve 250mA at <50V. Was there a '>' or a '<' in front of the warning?

I don't understand how a neutral earth fault would cause an RCD not to trip. The only reason I can think of for the fault on the socket circuit not to trip would be if it was sticky from lack of use (or a DC leakage issue).

If the cable isn't accessible and you cant install a new one, could you disconnect the faulty cable and derate it to 20A?

 

[Aaron b]

Sorry, I see the guy's above beat me to it, I don't know why I assumed a merger either... ?

 

[Roadhousezamora]

I was reading someones explanation and what I took from it was that the amount of current lost from the neutral was less than 30mA and therefore not triggering the RCD... Didn't fully understand how that happens though.

Yeah I may well have split the ring into two radials, there is a spare way in the board as well.

I have been testing on the circuit and not at the Rcd itself with no load so will try this when I'm next there.

If the Rcd responds at the device itself I will re add the loads but remove the neutral earth fault and see if they continue to operate.

I really appreciate this, being new my confidence when it comes to this sort of fault finding isn't great but I'm keen to make sure I get to the bottom of it, for myself as much as the client.

 

[pc1966]

What was the Zs at the test points used where the RCD failed? (Using the no-trip option, obviously!)

Seeing >50V warning to me sounds like the earth/CPC is exceptionally poor the test current is pushing it up above 50V.

 

[pc1966]

It's a Dilog 9110 (can't afford anything premium right now lol) it's ok but prone to being temperamental.

Have you tried a new set of probes?

I have the 9118 and it is fine (though I question the no-trip accuracy in some low Zs cases) and so far not seen the 50V warning. Probes are "consumables" as eventually the flexing breaks the conductors or the plugs get weak and don't establish as good connection.