View Single Post
  #11   Report Post  
Posted to uk.d-i-y
Robin Robin is offline
external usenet poster
 
Posts: 1,681
Default RCD protected socket.

On 02/11/2020 07:34, David Wade wrote:
On 01/11/2020 16:21, ARW wrote:
On 01/11/2020 11:27, Dave Plowman (News) wrote:
Have a two gang RCD protected metal clad socket in the conservatory
which
has failed - RCD will no longer latch on. It is part of one of the
kitchen
rings which isn't RCD protected at the CU - it feeds the washing machine
etc and other things which are left plugged in permanently.Â* Needs to be
an RCD as likely to be used for things outside the house.

Dunno the brand - it came from TLC. Notice you can pay between about £25
to £80 for similar. Is this sort of failure common at the cheaper end?


I would say so.

Called out a few months ago to the horse sales in Doncaster. This was
an unheated room next to the vets office. In order to get the 2m
distancing they decided to use this room.

Two 10 year old unused RCD metal clad sockets had failed.

Just a tip on buying a new one Dave. Try to get one to BS7288:2016 and
not old stock that is to BS7288:1990.


From some light googling it would appear that BS7288:2016 devices are
not suitable for use in this situation. The standard says they are only
to be used as additional protection where there is an additional RCD.


I don't have access to the full text but thought the wording was
different and more nuanced - viz "SRCDs are intended for use in circuits
where the fault protection and additional protection are already assured
upstream of the SRCD." So yes, an SRCD to BS 7288:2016 can /never/ meet
the requirements of BS 7671:2018 for additional protection. But then
nor would one to the 1990 standard. Luckily, Roland is not required to
bring the circuit up to the current regs. All I think he is doing is
making a like-for-like replacement (which is of course a darn sight
safer circuit than if he fitted a socket with no RCD protection). And
ISTM Adam is right that he might as well do so with one to the latest
standard.


--
Robin
reply-to address is (intended to be) valid