ProDave Problem, Lamp becomes live when CU switched Off!

[ProDave]

No, no loose wires. All busbars secure.

Unbeknown to me, the wire I was about to cut was live, in a round about way, before I started to cut it.

 

[NozSpark]

I'll give you another prompt to try and get you thinking.The main switch is turned off, so you would expect the whole house to be dead, but clearly it's not.

so might a good place to start be remove the cover from the CU and have a look at that main switch and how it's wired?

What would you like me to look for and what tests would you like me to do?

I could say something, but I won't:O

 

[samson337]

Where the tails in the rcd side

 

[ProDave]

No the tails are correctly fitted into the input side of the main switch.

Still nobody asking me to make any voltage measurements anywhere?

 

[sammers230]

Forgot to mention about testing between items and I suppose the first thing is I would check that the main switch is actually isolating the supply and if not then then start to turn circuits off in the board to ensure there are no circuits which are live back feeding the system.

 

[samson337]

Just another stab coz have not come across this before, had they connected the tail into an mcb from main switch, coz it don't say that you isolated the whole board unless I've missed it

 

[ProDave]

You are getting very warm with the main switch not isolating everything.

I'll let it run a little longer before posting the details of what was wrong.

 

[samson337]

Upside down main switch

 

[sammers230]

Could it be along the same lines as solar pv and having a dual supply but like a genny or something that for safety kicks in a set amount of time after the supply has been removed or had dropped? Either that although it's uncommon for domestic to have a genny or having an off peak circuit and after you switched off the board a misconnected circuit had backed fed the system once the off peak supply had switched on?

 

[ProDave]

No not a secondary supply, and no not the main switch upside down.

Think along the lines of what wiring fault would prevent the main switch from isolating everything. How can a circuit appear to be dead (i.e anything plugged in stops working) but it's not really dead?

 

[sammers230]

Well the looking dead but being live bit will almost certainly be reverse polarity but trying to work out how the main switch won't isolate everything because L-N reverse the main switch being double pole would still isolate no matter what way round they are and if L- E reverse then nothing would work in the first place.

 

[a12jpm]

Faulty main isolator only breaking neutral, circuits energise when they find a path back via earth when side cutters bridge them?

 

[batty]

I hope it hasn't got anything to do with Steps pet hate.

 

[philbas]

The neutral got broken the line didnt,

test betweeen all 3.

Why it happened i dont know

 

[ProDave]

I think that's close enough so here's what actually happened.

It was a split load CU with one side just from the main switch and the other side RCD protected.

The CU was either assembled wrong in the factory, or reconfigured by the installer and he got it wrong.

The problem was the L feed to the RCD was connected to the incoming side of the main switch. The neutral feed to the RCD was correctly wired to the output side of the main switch. So when the main switch was turned off, the circuits appeared to go dead but the line was still connected to the circuits on the RCD side.

Closing my cutters around the damaged cable, bridged E to N in that cable and "completed" the circuit.

See the attached diagram to show what happened and where the current flowed:

What lessons can we learn from this?

NEVER assume turning a switch off isolates everything.

ALWAYS test for dead.

In a case like this, even if you are not expecting to do any work inside the CU, take it's cover off to actually make sure ALL circuits are dead. You would have either visually spotted this fault, or have found the RCD side to still be live had you opened the CU and tested.

Even though I normally hate them, a quick swipe of the cable I was about to cut with a volt stick would have shown it to be live, so I would have gone looking for the fault before cutting it.

My old age and to some extend complacency made me wrongly assume everything was dead because the main switch was off. Don't follow my example, do it properly and test first.

And lastly, this shows whoever did the original install and original inspection and testing, failed big time. this was an obvious fault as soon as you opened the CU so it makes you wonder if the installation was ever tested properly.

 

[sparkytim]

So Post 11 was correct then

 

[ProDave]

So Post 11 was correct then

Yes, but you hadn't explained how part of the circuit still had it's line connection.

And I didn't want to reveal the answer so early on to give others a stab at what the fault was.

 

[springcrocus]

Sammy, in post 38, also got it right except he said tails instead of tail, which is what made all the difference.

 

[alextrician]

that's poor isolation procedure - you should make sure the whole installation is dead and probably the main switch was damage before and the current was arching through it...this is what i think of

 

[sparkytim]

that's poor isolation procedure - you should make sure the whole installation is dead and probably the main switch was damage before and the current was arching through it...this is what i think of

Suggest you try READING the posts before making a comment :shakehead ( maybe at least # 55 if you do not have time for all of it)