Outside light - tripping RCD

[hightower]

Went to a job last week, replace an outside light with another customer bought one. Straight swap, few extra holes in the wall for new mounting, job done. Photocell model.

Left the job with a working light. Now, I fitted this in the p***ing down rain, which in itself was a tad silly of me but I wanted this job out the way. I chalked it off as something damp inside and said I'd be back round to look in a few days when it dried up - not a problem as it has a switch so can be isolated leaving the rest of the electrics on and working.

When I went back it hadn't rained since fitting, so should have been nice and dry in there by now. I checked and there was no wires trapped or anything. I'd asked if they'd tried it since and "no", so flicked it on and it works. Left the job, 10 minutes later a message to say it's tripping again.

So the symptoms are the light works fine for a little while, but then starts tripping. Once it starts tripping it continues tripping until seemingly left a while.

Probably a simple one as "dodgy fitting" but wanted to double check on here if anyone had any more thoughts? He's asked me to stick his old one back up, which I'll be doing this weekend at some point, but wanted a second opinion....

 

[sparksburnout]

I'd still of IR tested the install....would take 5 minutes to do a global. As it is you've made an assumption that may be correct but it's still an assumption.

How do you mean 5 minutes to do a global? Usually takes about 20 mins for the occupier to run round and unplug everything. Then you have to spend another 10+ mins working out whats still connected......problems pretty obvious here why complicate it??

 

[westward10]

How do you mean 5 minutes to do a global? Usually takes about 20 mins for the occupier to run round and unplug everything. Then you have to spend another 10+ mins working out whats still connected......problems pretty obvious here why complicate it??

Nah leave it all plugged in, IR to earth job done. Does not more harm than a PAT, if you get problems then unplug.

 

[sparksburnout]

Nah leave it all plugged in, IR to earth job done. Does not more harm than a PAT, if you get problems then unplug.

Yeah right on.

 

[sparksburnout]

I didn't explicitly say I didn't test the install. I did IR test it and returned fine. Also clamp tested to see but that returned nothing interesting

Nice one, don't make a drama out of it. Problems obvious, no need to conduct an inquest, times money! Tricky concept for some folks.....

 

[Des 56]

Nah leave it all plugged in, IR to earth job done. Does not more harm than a PAT, if you get problems then unplug.

Yet another who deviates from the ABC of testing to improve and speed up fault finding in Domestic installs
After all,it's the fault that needs to be located and that can be installation or user cause,two birds with one stone and all that

 

[Davisonp]

Do you know the trip current for the main RCD (Ramp test)? And the current board leakage, I have had a couple of houses right on the limit for the RCD, then adding something with acceptable leakage pushes it over the limit.

 

[Deleted member 9648]

How do you mean 5 minutes to do a global? Usually takes about 20 mins for the occupier to run round and unplug everything. Then you have to spend another 10+ mins working out whats still connected......problems pretty obvious here why complicate it??

Nah leave it all plugged in, IR to earth job done. Does not more harm than a PAT, if you get problems then unplug.

Nice one, don't make a drama out of it. Problems obvious, no need to conduct an inquest, times money! Tricky concept for some folks.....

5 minutes! As above,you are looking for a fault not doing initial verification,so leave everything plugged in and you are more likely to locate the problem.
Over the years I've been to loads of jobs where a previous sparks has disconnected an 'obvious problem' only for the tripping to recur. 5 minutes testing found an issue on a completely unrelated circuit.
Your cavalier approach is a recipe for comeback.

 

[sparksburnout]

5 minutes! As above,you are looking for a fault not doing initial verification,so leave everything plugged in and you are more likely to locate the problem.
Over the years I've been to loads of jobs where a previous sparks has disconnected an 'obvious problem' only for the tripping to recur. 5 minutes testing found an issue on a completely unrelated circuit.
Your cavalier approach is a recipe for comeback.

That's me, specialised subject the bl**ding obvious, as Basil would say.