Three layers of RCD protection

[RandyBender]

Hello,

I want to add an extra socket in my garage as I only have one existing. The only thing that's puzzling me a bit is that the one that is there is one of those metaclad sockets with RCD built in.
This connects to the garage CU that has 16A for the socket, 6A for the light and an RCD protecting both.
This garage CU connects back to the house via some SWA cable and is on the RCD side of the board there too.

I'm a bit curious about the set-up. I guess the RCD in the house protects the cable run to the garage. The RCD in the garage possibly serves no purpose but is it good practice to have another more local to socket? And the RCD built into the socket serves no purpose at all, is that right?

And the purpose of this is just to confirm that I'm safe to buy a standard socket with no RCD to connect to the one that's there?

Thanks for reading.

 

[ban-all-sheds]

There's no need to have an RCD socket.

With RCDs in series it'd be pot luck which one would trip anyway.

What you should have is just the RCD in the garage CU.

 

[rebuke]

Indeed, RCDs in series is generally pointless (and implies it wasn't properly installed, as there is no way the RCDs beyond the first one could have been properly tested, since the first one may have tripped anyway!)

The only case where it isn't is if you have a higher rating RCD or a time delay type fitted upstream, so for example a TT installation might have a 100mA RCD protecting everything, but you'd then want a 30ma one at a suitable point to protect the sockets...

 

[ban-all-sheds]

for example a TT installation might have a 100mA RCD protecting everything, but you'd then want a 30ma one at a suitable point to protect the sockets...

Old skool....

 

[Risteard]

The only case where it isn't is if you have a higher rating RCD or a time delay type fitted upstream

Just time-delayed or S-type RCD. 100mA RCD won't provide discrimination (selectivity) with 30mA device if it isn't time-delayed.

 

[ericmark]

RCD's are fitted both ends of the cable with caravan supplies but that is an odd case.
Normally it goes 3x rule so if supply is 100ma and then ring main 30ma and socket for lawn mower 10ma.
However active RCD's are also used to protect where motors are used so if you plug in something then realise no power and return to house to turn back on the item will not start until reset button is pressed.
My grinder has a RCD plug so it will comply with auto restart rules.
Pond has 10ma so it will not trip rest of house.

 

[Adam_151]

, as there is no way the RCDs beyond the first one could have been properly tested, since the first one may have tripped anyway!)

I never have a problem testing them properly individually

Often see two rcds in series where pre-fab site cabins are connected up to a site supply, the cabins are supplied with an RCD incommer DB and the same is installed at the orgin of the supply as its TT.

(the trick is to put the earth lead from your rcd tester on the neutral conductor before the RCD rather than the earth

 

[securespark]

I never have a problem testing them properly individually

I always have a problem testing them properly individually.

How can you not? Unless you disconnect the outgoing conductors?

 

[Adam_151]

You might want to put your glasses on and look at my post again, simon

 

[SimonH2]

The only case where it isn't is if you have a higher rating RCD or a time delay type fitted upstream

Just time-delayed or S-type RCD. 100mA RCD won't provide discrimination (selectivity) with 30mA device if it isn't time-delayed.

Correct - something else a lot of people don't realise.

For those thinking a simple current sensitivity will provide discrimination think about this. You have a fault such a person toughing a live wire, the fault current rises rapidly. Your 30ms RCD will trip - but by the time it does the fault current sensed by the upstream RCD will also have exceeded it's tripping point. So they should ALL trip together.

By using time discrimination, the 30ms RCD trips, but the 100ms RCD doesn't because the fault is removed before it's trigger time is reached.