RCD trip by non-RCD protected side of circuit

[James_H]

Sorry for yet another 'my RCD has tripped' thread, but I have not found this through searching.

Since having by electrics upgraded a couple of months ago to include RCD protection on the sockets side, I have had a couple of rogue trips. It could have been the dishwasher (new) or tumble drier (older) either time before.

However, last night it went off after a quick on-off of the fluorescent strip light in the garage. The garage is integral, under a bedroom.

I have a TNCS arrangement, with separate consumer units for the RCD side (socket circuits except fridge freezer) and non-RCD side (everything else, including the garage light mentioned). The division is made via a Henley just after the meter.

Should it be possible for a circuit on the non-RCD side to trip the RCD?

Is this normal or does it suggest something is amiss? I cannot see how any rogue leakage to earth from a lighting circuit could affect the RCD side.

 

[sparky40]

Hi,Can you post a pic of the consumer units at all ? It sounds like the RCD is covering both boards at present.

 

[RF Lighting]

Have a read of this. It may not be quite the same but may give you a few ideas to get you started

 

[James_H]

Only after posting I realised I had the wrong spelling of fluorescent in one of my searches! And I somehow missed that other recent similar problem.

Due to space constraints there was no room for a split load board, so one CU covers the non-RCD circuits and another CU the RCD-required circuits. There is no single point of isolation (NIC spark should really have put a sticker on to say so), with the RCD also being the main switch for that CU.

The only point where the RCD and non-RCD sides meet (in theory) is a henley just after the meter and the main earthing block.

The other post stated that a fluorescent light may cause backfeeds, but how does it trip a separate consumer unit?

Another question: the BS7671 test certificate only covered the new circuits added to the existing non-RCD CU, plus all circuits on the new CU. Should the existing unaltered (lighting) circuits have also been tested?

 

[WSDLTD]

cross conected neutral on split load board

 

[securespark]

back feeds and transients from flu's and pc's can be to blame.