Downstream RCD

[MegOmz]

Hi All
Might be an easy answer for most but I am just getting back in the game after 20 years...

I am time served and did my 18th last year, I have a question concerning selectivity. Here is the scenario. A domestic property has an 18th edition dual RCD board and one of the final circuits is providing a sub main to DB2 which is for an extension ( 40A 30898 MCB and 10mm T&E)
This DB is only 5 metres away in The utility room and is serving the new extension. DB2 is also RCD protected. If a fault occurs on a final circuit On DB2 It’s pot luck as to which RCD trips, DB1 or DB2.
what is the preferred (and BS7671 compliant) way if you were to swap out DB2? New CU with just MCB’s as it’s already protected by an upstream RCD? i don’t think you could move the mcb in DB1 to being unprotected as the sub main is buried and is T&E. I believe you would have the same issue even if you put RCBO’s in DB2, also if you were to perform an eicr what RCD trip times would you put on the sheet as three tests could trip DB1 and and two would be for DB2 so not really ideal
also don’t think you could use an S type in DB1 due to the disconnection times required.

thanks for reading !

 

[westward10]

You cannot use a time delayed device (30ma do not exist in the UK) as the 10.0 requires additional protection. One option is to ditch any rcd protection at DB2 this removes the rcd selectivity issue. If I were doing an EICR I would Code 3 it as it isn't really a safety issue, it is was it is.
As for recording/testing the rcds I would take the upstream device out of the question to test the downstream device but that is just me and not something I would necessarily demonstrate.

 

[MegOmz]

You cannot use a time delayed device (30ma do not exist in the UK) as the 10.0 requires additional protection. One option is to ditch any rcd protection at DB2 this removes the rcd selectivity issue. If I were doing an EICR I would Code 3 it as it isn't really a safety issue, it is was it is.
As for recording/testing the rcds I would take the upstream device out of the question to test the downstream device but that is just me and not something I would necessarily demonstrate.

Thanks for the quick reply - totally agree with the S type. my personal option would be to just have MCB’s in DB2 but just looking for clarification it’s OK. I can’t see why not and I have scoured BS7671 nothing specifically mentioned. as for the eicr I can see exactly where you are coming from, makes complete sense
thanks again

 

[Lucien Nunes]

If it were not for the need to provide additional protection to the submain, there would be little difference between having the RCD protection in DB1 versus using an unprotected way in DB1 and a single RCD in DB2 as the main switch.

But if DB2 serves both lights and sockets in the extension, good practice would be to split those circuits across two RCDs or RCBOs so that one circuit can remain operational in the event of a fault on the other, just as you would split them across the two RCDs in DB1 if they originated there.

Clearly the need to protect the submain scuppers this, and possibly the lack of non-RCD ways in DB1 even if the submain were SWA. In this situation, stuck with an RCD in DB1, I would definitely avoid putting another in DB2.

 

[MegOmz]

If it were not for the need to provide additional protection to the submain, there would be little difference between having the RCD protection in DB1 versus using an unprotected way in DB1 and a single RCD in DB2 as the main switch.

But if DB2 serves both lights and sockets in the extension, good practice would be to split those circuits across two RCDs or RCBOs so that one circuit can remain operational in the event of a fault on the other, just as you would split them across the two RCDs in DB1 if they originated there.

Clearly the need to protect the submain scuppers this, and possibly the lack of non-RCD ways in DB1 even if the submain were SWA. In this situation, stuck with an RCD in DB1, I would definitely avoid putting another in DB2.

Thanks Lucien, much appreciated

 

[westward10]

What is it you are doing at this site.

 

[MegOmz]

What is it you are doing at this site.

Sorry Chap which site ?

 

[westward10]

Sorry Chap which site ?

Sorry misread your first post where it says scenario.

 

[MegOmz]

Sorry misread your first post where it says scenario.

 

[MegOmz]

This was my first post short of introducing myself. I have wanted to ask it for a while. I belong to a few electrical groups on Facebook and based on the responses I have seen for other questions I didn’t want to set myself up! This forum seems to be for people who actually want to help without ridicule.
having been out for a while It’s been interesting to see the changes. thanks again

 

[catdog1121]

I would either change the T+E to SWA so you can just use a non RCD protected mcb in DB1 or put DB2 as an rcbo in DB1 and swap the RCD with a main switch in DB2.

Just put a label on DB2 saying RCD protection provided at DB1, in the even of loss of power check at DB1.

 

[moose]

Could you say....

Isolator/RCBO40amp(Feed DB2)/RCD(circuits)/RCD(Circuits) if you have space in the board.

Have fun.

 

[Strima]

Does the installation method for the 10mm T&E for the distribution circuit require additional protection?

 

[pc1966]

Does the installation method for the 10mm T&E for the distribution circuit require additional protection?

Good point, as RCD it is not needed because it is T&E, but because it is hidden but vunerable to DIY accidents.

 

[Lucien Nunes]

He did say:

i don’t think you could move the mcb in DB1 to being unprotected as the sub main is buried and is T&E.

in which I took 'buried' to mean possibly at a depth requiring additional protection. But it is a good point that this was not confirmed as a fact.

 

[MegOmz]

Does the installation method for the 10mm T&E for the distribution circuit require additional protection?

Hi, it would I believe as is buried less than 50mm, thanks

 

[MegOmz]

Could you say....

Isolator/RCBO40amp(Feed DB2)/RCD(circuits)/RCD(Circuits) if you have space in the board.

Have fun.

Hi Moose
if You put RCDs in DB2 would you have the same issue with tripping? it could trip the RCBO in DB1 or the RCDs in DB2.
also DB1 is already a split load board with two RCD,s. or have I miss understood? Sorry if I have
thanks

 

[moose]

I was Just thinking if you placed DB2 in the utility room on its own circuit within DB1 moving the new RCBO 40amp to feed new DB2 next to the Isolator. Not apart of the RCD split board circuits.

Isolator - Feeds new RCBO. and x2 RCDs within DB1

I am assuming the protection of the Feeder cable to DB2 is by BD1. Will only trip Extension.
I am assuming yes it will trip off all circuits in Utility room. Unknown x Circuits?
I am assuming Fault Future, be solved by circuit elimination to get power up and running.
My assumption, Home owners do this with split boards Left side or Right side now. Then get a qualified man like yourself to fix the issue.

Others can quote regs for better fun to the story so far.

Have fun.

 

[Simon47]

It’s pot luck as to which RCD trips, DB1 or DB2.

Or quite likely, both of them.
Unless they have significantly different tripping times, the slower device is likely already to be "in motion" on it's way to tripping by the time the faster one actually opens the circuit and removes the trigger current. I don't know what the time lag is between "irretrievably on it's way to opening" and "contacts opened" for a typical RCD, but I'd imagine it's a significant part of the tripping time.

 

[pc1966]

If you have a DB fed off another DB using a MCB, RCBO, or RCD/MCB combination you will have poor selectivity.

The best you could hope for in this situation (vulnerable 10mm T&E) is to have the supply off, say a 45A C-curve RCBO and ditch the RCD in the sub-DB. But it is still going to be poor selectivity, due to several circuits sharing the same RCD trip, and that down-stream MCBs will probably also trip the RCBO on over-current fuse-blowing faults (PSSC above 225A might trip the feed RCBO).

Now that is not dangerous, just poor design and inconvenient.

If you really want to fix it, you would need SWA sub-main off a fused-switch and split the tails feeding the main DB. Then you might get away with a 50A fuse or so, and have a sporting chance of selectivity with down-stream MCB or RCBOs. But the cost & time of doing that might be too much.

In any case, for a new year, and hopefully for a better year!