UK RCD issue baffling me

[ChiliTom]

I have a dual board consumer unit. Both sides have a RCD. The left hand side of the board has all the MCBs for all the 6 ring mains for the property. The right side of the board has the lighting and hob and cooker for the house.

Whenever a current is drawn from a socket the right side RCD trips, but not the left side. So my lights trip but sockets stay on. It happens on any of the 6 ring mains. And seems to be any appliance that draws a large current such as a hoover, kettle, immersion heater etc.

I do have a alarm system but I have totally isolated it by unwiring it front the spur it was connected to. Still the same issue.

I do also have under cabinet lights that are ran from a ring main... I have yet to investigate these.

The MCBs switched off were because they were just straight up tripping the right side RCD. Probably because they all had stuff like a fridge, boiler etc drawing current. The right hand side lighting circuit that is turned off also automatically trips the right RCD.

Electrician has tested each circuit and can't find a fault. The board has been wired correctly with no crossed over neutrals.

Anyone got any clue where to start some more fault finding? I am a ex Army Electrician, that was 17th ed trained... But that was about a decade ago but I know enough to be safe.

Cheers

 

[brianmoooore]

Is it the same electrician who installed the CU that's doing the testing now? Not exactly impressed with the distribution of circuits between the two RCDs in that CU.
It'll be an earth to neutral fault in one of the RH bank circuits, and things with ceramic, metal enclosed elements don't always show poor IR when cold.

 

[mainline]

The MCBs switched off were because they were just straight up tripping the right side RCD. Probably because they all had stuff like a fridge, boiler etc drawing current. The right hand side lighting circuit that is turned off also automatically trips the right RCD.

Turning off single pole mcbs won't disconnect the N to Earth fault, best to disconnect individual circuits entirely.
If it trips instantly with all circuits energised, then it shouldn't take long to isolate the faulty circuit.

 

[ChiliTom]

Is it the same electrician who installed the CU that's doing the testing now? Not exactly impressed with the distribution of circuits between the two RCDs in that CU.
It'll be an earth to neutral fault in one of the RH bank circuits, and things with ceramic, metal enclosed elements don't always show poor IR when cold.

Why would the RH bank of lighting circuits RCD trips though from using plus sockets that are on the left bank.

 

[ChiliTom]

Is it the same electrician who installed the CU that's doing the testing now? Not exactly impressed with the distribution of circuits between the two RCDs in that CU.
It'll be an earth to neutral fault in one of the RH bank circuits, and things with ceramic, metal enclosed elements don't always show poor IR when cold.

No it's a different Electrician. We both actually commented on the split of the circuits as being sub par. But it shouldn't affect the RCDs as it is.

 

[brianmoooore]

Because most of the neutral current from the LH circuits will flow back the neutral tail to the meter and then the service head, but a small proportion will pass through the RH RCD neutral side, down the neutral wire of the faulty circuit, through the neutral - earth fault, back the cpc of the faulty circuit to the CU, then down the cpc to the service head, where, in the case of TNC-S, it is joined to the neutral. If this current momentarily reaches about 20mA (typical tripping current of a 30mA RCD), then the RH RCD will trip.

 

[ChiliTom]

Because most of the neutral current from the LH circuits will flow back the neutral tail to the meter and then the service head, but a small proportion will pass through the RH RCD neutral side, down the neutral wire of the faulty circuit, through the neutral - earth fault, back the cpc of the faulty circuit to the CU, then down the cpc to the service head, where, in the case of TNC-S, it is joined to the neutral. If this current momentarily reaches about 20mA (typical tripping current of a 30mA RCD), then the RH RCD will trip.

So the left hand bank that has all the sockets on it, isn't to fault, even though it trips the other sides RCD. The fault lays somewhere in one of the left hand lighting circuits and will be a E/N fault?

 

[brianmoooore]

.......... the left hand lighting circuits ..........

Typo?

 

[ChiliTom]

Typo?

Yes sorry right hand lighting circuits

 

[brianmoooore]

N - E fault will be in the RH circuits, and could be a lighting circuit, especially if any of the lights are outdoors, but I'd concentrate on the two circuits that are labelled as cookers/ovens/hobs first of all, even if it's just turning off their DP isolators when not in use.

 

[ChiliTom]

I still don't understand how a LH bank of fully normal and fully operational set of 6 ring mains trip the opposite sides RCD when they draw current. Surely the neutral current goes back to the neutral bar then back to the mains. It never passes through the other RCD.

 

[brianmoooore]

As I explained earlier, there is a parallel path for the neutral current to return to the service head. One that results in a small proportion of the neutral current from the socket circuits, but not the live current, flowing through the RH RCD, thus unbalancing it.
The proportion of diverted current is the ratio of the resistance of the neutral tail compared to the resistance of the RH circuit neutral, fault resistance, RH circuit cpc, and main cpc, which can be a very small proportion, BUT 20mA is a very small current, when compared to the many tens of amps that can flow as an inrush current on the socket circuits.

 

[ChiliTom]

As I explained earlier, there is a parallel path for the neutral current to return to the service head. One that results in a small proportion of the neutral current from the socket circuits, but not the live current, flowing through the RH RCD, thus unbalancing it.
The proportion of diverted current is the ratio of the resistance of the neutral tail compared to the resistance of the RH circuit neutral, fault resistance, RH circuit cpc, and main cpc, which can be a very small proportion, BUT 20mA is a very small current, when compared to the many tens of amps that can flow as an inrush current on the socket circuits.

Ok so I should disconnect the neutrals from the RH side one neutral bar by one and see which one once removed doesnt trip the RCD? This should indicate which circuit the fault lies in?

Then I guess just investigate the circuit and see if it's wired correctly?

 

[ChiliTom]

So am I right in saying when I use a socket for a kettle the current is going from the mains through the LH RCD to the MCB to the kettle and the return current is then going down the neutral back through the LH RCD (all ok at this point) up to the neutral bar where the RH RCD neutral is also connected, but because there is a N to E fault in a circuit that side the current is instead flowing down the RH RCD neutral to the earth somewhere which then trips the RCD as there was a imbalance (due to no live current being drawn that side)

 

[brianmoooore]

If you just disconnect the neutrals, leaving the lives connected, you stand a good chance of raising the voltage on those neutrals, so that the N-E fault now becomes a L-E fault, and now thrips the RCD because of the imbalance resulting from that.
IR testing with a MFT is the usual way to track down a fault like this.