2 x Main 2P RCDs in Split-phase DB: Weird tripping behaviour.

[Mark42]

Two phases supplying two 2-pole RCDs in a consumer unit reconfigured as a small DB (with an external 3-pole main switch). Common Neutral.

In an outbuilding supplied by 10mm 4-core. TN-C-S ‘exported’ from main house by 10mm bonding. Local TT as an option in progress, awaiting tests before deciding.

L1 has 2 x 400W Metal Halide lamps through a 10A C-curve breaker, L2 ditto but 3 x MH’s. Plus other finals not finished yet.

One, or sometimes both, RCDs immediately trip when both phases are energized, and BOTH the MH lights switches (in the same grid switch box) are switched on. Only one circuit on, (L1 or L2) never any problem.

Disconnecting either phase supply from its RCD, then shorting the L1/L2 bus bars to run the whole panel (and all the lights) from only one phase and one RCD, the problem disappears. It’s fine whether L1 or L2 is used as the supply.

I’m stumped. I’m asking on here because it may be obvious to someone more experienced than me. Have I missed something fundamental?

Thanks, Mark.

 

[johnboy6083]

have you eliminated the circuits being faulty? IE, carried out IR tests?

 

[Mark42]

have you eliminated the circuits being faulty? IE, carried out IR tests?

Only the cable, which is new, in a new building. IR was off the scale at 1000Vdc.

But the circuits now have ballasts and ignitors connected inside the MH fittings, which are up on the roof and inaccessible now the hired scissor lift has gone.

I was hoping not to have to mess with them for a long time. As I said each circuit works fine on its own, or together when supplied from only one RCD and one phase.

 

[Sintra]

Borrowed neutral??

 

[mechelec]

I would say it is the common neutral that is the problem. Can you separate the neutrals and try it. You need to keep the neutral for each rcd separate and basically treat it as split load board just with each rcd supplied by a separate phase.

 

[Mark42]

Sintra and Mechelec: Thanks for your perceptive answers!

I though about the neutrals but couldn’t understand why it would make any difference electrically, so ignored it.

Currently it’s:

L1, L2, N into external ‘Main Switch’

2 x 10mm T&E from here to the DB, ie two neutrals in parallel, terminated together in the main N bar in the box. (I used T&E for DI protection behind the panel, I didn’t bother with conduit and singles). This parallel arrangement doubles the Neutral CSA for a short run, is this the problem?

From the main N rail two separate N leads to the 2P RCDs.

Then from the switched side of the RCDs, two neutral wires to two separate (unconnected) N bars for the final circuits.

I’ll bodge the run from the main switch directly to the two RCDs in singles today, ie eliminating the single common neutral bar in the DB, and report.

I can’t see how it’s electrically any different, but maybe I’m missing something!

 

[johnboy6083]

I completely missed the fact that the neutral wasn't split inside the DB.

 

[RogertheBodger]

Are you sure that the Neutrals from your two circuits are going to the correct Rcd? if they are crossed, that could give you the symptoms you describe.

 

[Mark42]

Good point Roger! That could be it. Once it's a bit warmer here in Norfolk I'll go out to check.

Your nickname made me laugh! We should be 'friends' on here

 

[RogertheBodger]

Hope so, had to get my pencil out for that one.

If it doesn't work, could you show us a picture?

 

[Mark42]

A light bulb has just gone on!!

Of course! If the neutrals are 'shared' inside the DB some of the return current goes the 'wrong' way so the RCD/s think there's current imbalance!

Is that it?

I was being thick. What a rubbish install! Obvious once you realise?

Bloody hell, pencil out? The day-rate has just doubled

 

[RogertheBodger]

Not sure what you mean by 'shared' in this case. You have said that from the switched side of the RCDs, two neutral wires go to two separate (unconnected) N bars for the final circuits, which sounds OK to me. But you're on the right track.

 

[Mark42]

Not sure what you mean by 'shared' in this case. You have said that from the switched side of the RCDs, two neutral wires go to two separate (unconnected) N bars for the final circuits, which sounds OK to me. But you're on the right track.

Shared on the supply side. Like I said, I can't understand how it makes a difference. Perhaps they are crossed on the RCD side as you suggested, and I didn't notice it: it is difficult to QC your own work, you keep looking at the same mistake and just don't see it.

I'd better get out of the warm kitchen and down to the building.

My day-rate’s doubled too: It’s now TWO packets of peanuts and FOUR bananas!

 

[johnboy6083]

i bet when you find the problem youll go ape!

 

[Mark42]

Sorted. A totally stupid mistake which I just didn’t think about. I was too transfixed on the DB.

For some unknown reason I’d forgotten there was a two-phase supply to the light switches in the top 3-gang grid enclosure (see pictures.)

I’d commoned all the neutral wires and forgotten about it. (I didn’t common the phases though, that would have been easier to find … )

Oh well, live and learn. I can’t explain it. Thanks for all the advice.

It all makes sense now. When both DP RCDs and at least one final circuit is on, a proportion of the neutral return goes back through the ‘wrong’ RCD. Current imbalance sensed on one or, if it’s quick enough both, RCDs. It completely explains the erratic behavior.

Does anyone else remember the old ROSPA poster below? (Pre-dates the HSE). I’ve had it for about 25 years. In this case that IS the electrician

 

[RogertheBodger]

Johnboy warned you about all those Peanuts and Bananas!