Manator
©Honorary Essex Boy™
Think why the circuit was not completed until prodave used his cutters
Another "impossible" fault thread
- Thread starter ProDave
- Start date
Help Support Talk Electrician Forum:
Rcd feeding one half of cu connected to wrong side of main switch??
Neutrals in wrong places/bars?
- Joined
- Nov 28, 2009
- Messages
- 14,762
- Reaction score
- 951
So why did the lead lamp go off when I turned the main switch off?
Manator
©Honorary Essex Boy™
Prodave, when you give the answer will you describe how you found it, I think a couple of guys would like to know the test procedure. Very good exercise and an excellent example.
- Joined
- Nov 28, 2009
- Messages
- 14,762
- Reaction score
- 951
Okay it's answer time.
Firstly a note about safe isolation. Once you read what the fault was, you need to learn from it, and possibly update your idea of "safe isolation" and in particular "test for dead"
Having been in this game for a while, I have perhaps become complacent. When isolating one circuit to work on, while leaving others energised (for instance I want to work on a socket circuit, but it would be nice to leave the lights on so I can see what I am doing) then I ALWAYS test for dead on the circuit I am going to work on.
But if, as in this case, I have "isolated" the whole house buy turning off the main switch, then I rarely test for dead, as it can't possibly be live can it. Can it?
WRONG. IT CAN BE LIVE AS THIS EXAMPLE IS ABOUT TO SHOW, SO ALWAYS ALWAYS TEST FOR DEAD.
Once I found the fault, my first check was plug my socket tester into the socket that my lead lamp was in. It showed "L-E reversed" but if I then unplugged the lead lamp, it then showed "No Neutral"
The fault was indeed "no neutral" and the "L-E fault was a bit like the polarity reversal fault a few days ago in another thread.
But we would expect "no neutral" because the isolator was off. So really the fault was why was there still a Live connection with the isolator off?
To make any meaningful measurements, it was off with the CU cover. The fault was almost immediately obvious.The split load kit was installed incorrectly. Whether this was a factory fault, or was done by the installer I don't know, but it SHOULD have been picked up by whoever inspected and tested the original install, but it wasn't. It should also have been picked up if the property ever had a PIR, but being only 4 years old, I doubt it's ever been inspected.
As I was only there to fix a simple damaged cable, I would not have thought to test for such things, or even look inside the CU. But for the strange happenings with my lead lamp, and having found a fault, I would have been negligent to ignore it.
So what had happened. Well simply the cables feeding the RCD were wrong. The neutral in to the RCD came from the first neutral busbar, so that was okay, when the main isolator was off, the neutral to all circuits was isolated. BUT the live in to the RCD was connected to the INPUT side of the main isolator (squeezed in with the meter tail), so with the isolator off, the RCD still had a Live feed, but no neutral. And of course I didn't turn the RCD off, because isolating the main switch kills all the power. WRONG ASSUMPTION.
So my lead lamp still had L connected. So the isolated neutral would have been floating up near 240V fed via the filament of my lamp. So if you grounded any neutral anywhere, then the lamp would light.
But it was a brown wire in the 3 core link between upstairs and downstairs landing light switches that I grounded. How does that work, that's not a neutral is it?
Well that's down to a VERY poor choice of cable colours. See my rather poor sketch attached (the red indicating the current path that lights the lamp).
The landing light is wired in a configuration I often use myself, where a L feed goes into the downstairs switch, then you send Neutral and two Live strappers up the 3 core. Resulting in both the L and N for the landing light conveyed by the 3 core cable, thus avoiding any possibility of borrowed neutral. Now when I do that, I use the black core as neutral, and the brown and grey as the two live strappers. Well this spark, in his infinite wisdom had used BROWN as neutral, and black and grey as the two Live strappers. That's another subject for debate, but that is NOT a colour scheme that I would use.
So there is the fault, a L supply not isolated, a floating neutral, and a BROWN neutral shorting to earth energised my lead lamp.
So what lessons can we learn?
ALWAYS test for dead. Even when you think nothing can possibly be live. But how would you test for dead here? I'm not altering a switch or other accessory (where it would be easy to properly test for dead with a meter because you have access to all the cores where they are terminated). Instead I'm just cutting a physically damaged cable to repair it. So how can I "test for dead" a cable before I cut it? Well the cable has been damaged, and two conductors are showing (as it happens brown and CPC) so I could try and get my meter probes on there, but I'm up in the loft with nothing else near by to give an earth reference. Or I could use my neon screwdriver. I actually contest the neon would be the best bit of kit here, and certainly that's what I usually use if there's any doubt. But I was so certain it was isolated, I didn't even bother using that.
No harm came of course as I was using an insulated pair of cutters. Obviously (I hope) we all do, but let this warn you why you use insulated tools.
Anyway, it made my afternoon a little more interesting. This is the first time I have encountered a CU where the main switch did not properly isolate everything, so you can teach an old dog new tricks.
Perhaps now, even when isolating everything at the main switch, then opening up the CU to check the main switch is wired correctly is the only safe way, and actually measure that nothing is coming out of the main switch.
And finally, just because everything goes off when you turn the main switch off, do not assume that means it is safe to work on.
Firstly a note about safe isolation. Once you read what the fault was, you need to learn from it, and possibly update your idea of "safe isolation" and in particular "test for dead"
Having been in this game for a while, I have perhaps become complacent. When isolating one circuit to work on, while leaving others energised (for instance I want to work on a socket circuit, but it would be nice to leave the lights on so I can see what I am doing) then I ALWAYS test for dead on the circuit I am going to work on.
But if, as in this case, I have "isolated" the whole house buy turning off the main switch, then I rarely test for dead, as it can't possibly be live can it. Can it?
WRONG. IT CAN BE LIVE AS THIS EXAMPLE IS ABOUT TO SHOW, SO ALWAYS ALWAYS TEST FOR DEAD.
Once I found the fault, my first check was plug my socket tester into the socket that my lead lamp was in. It showed "L-E reversed" but if I then unplugged the lead lamp, it then showed "No Neutral"
The fault was indeed "no neutral" and the "L-E fault was a bit like the polarity reversal fault a few days ago in another thread.
But we would expect "no neutral" because the isolator was off. So really the fault was why was there still a Live connection with the isolator off?
To make any meaningful measurements, it was off with the CU cover. The fault was almost immediately obvious.The split load kit was installed incorrectly. Whether this was a factory fault, or was done by the installer I don't know, but it SHOULD have been picked up by whoever inspected and tested the original install, but it wasn't. It should also have been picked up if the property ever had a PIR, but being only 4 years old, I doubt it's ever been inspected.
As I was only there to fix a simple damaged cable, I would not have thought to test for such things, or even look inside the CU. But for the strange happenings with my lead lamp, and having found a fault, I would have been negligent to ignore it.
So what had happened. Well simply the cables feeding the RCD were wrong. The neutral in to the RCD came from the first neutral busbar, so that was okay, when the main isolator was off, the neutral to all circuits was isolated. BUT the live in to the RCD was connected to the INPUT side of the main isolator (squeezed in with the meter tail), so with the isolator off, the RCD still had a Live feed, but no neutral. And of course I didn't turn the RCD off, because isolating the main switch kills all the power. WRONG ASSUMPTION.
So my lead lamp still had L connected. So the isolated neutral would have been floating up near 240V fed via the filament of my lamp. So if you grounded any neutral anywhere, then the lamp would light.
But it was a brown wire in the 3 core link between upstairs and downstairs landing light switches that I grounded. How does that work, that's not a neutral is it?
Well that's down to a VERY poor choice of cable colours. See my rather poor sketch attached (the red indicating the current path that lights the lamp).
The landing light is wired in a configuration I often use myself, where a L feed goes into the downstairs switch, then you send Neutral and two Live strappers up the 3 core. Resulting in both the L and N for the landing light conveyed by the 3 core cable, thus avoiding any possibility of borrowed neutral. Now when I do that, I use the black core as neutral, and the brown and grey as the two live strappers. Well this spark, in his infinite wisdom had used BROWN as neutral, and black and grey as the two Live strappers. That's another subject for debate, but that is NOT a colour scheme that I would use.
So there is the fault, a L supply not isolated, a floating neutral, and a BROWN neutral shorting to earth energised my lead lamp.
So what lessons can we learn?
ALWAYS test for dead. Even when you think nothing can possibly be live. But how would you test for dead here? I'm not altering a switch or other accessory (where it would be easy to properly test for dead with a meter because you have access to all the cores where they are terminated). Instead I'm just cutting a physically damaged cable to repair it. So how can I "test for dead" a cable before I cut it? Well the cable has been damaged, and two conductors are showing (as it happens brown and CPC) so I could try and get my meter probes on there, but I'm up in the loft with nothing else near by to give an earth reference. Or I could use my neon screwdriver. I actually contest the neon would be the best bit of kit here, and certainly that's what I usually use if there's any doubt. But I was so certain it was isolated, I didn't even bother using that.
No harm came of course as I was using an insulated pair of cutters. Obviously (I hope) we all do, but let this warn you why you use insulated tools.
Anyway, it made my afternoon a little more interesting. This is the first time I have encountered a CU where the main switch did not properly isolate everything, so you can teach an old dog new tricks.
Perhaps now, even when isolating everything at the main switch, then opening up the CU to check the main switch is wired correctly is the only safe way, and actually measure that nothing is coming out of the main switch.
And finally, just because everything goes off when you turn the main switch off, do not assume that means it is safe to work on.
Well done Dave, but I thought you said earlier that the Cu was wired correctly.
I hate these as I can never get my head around them, like Steps I don't do lapdancing.
I kept reading and came to the conclusion of the rcd fed from the L side of the MS somehow, but not in as much detail as the actual fault required.
Nor did I even consier the colour coding.
Well done.
I have wired split boards wrong myself when in compromising situations, read awkward, where the CU HAS to go where it was and you are on the top of the ladder in a hole with little light and p1553d off.
However, I have managed to find it in time, normally when one of the rcd's won't stay in!
Thanks Dave.
I hate these as I can never get my head around them, like Steps I don't do lapdancing.
I kept reading and came to the conclusion of the rcd fed from the L side of the MS somehow, but not in as much detail as the actual fault required.
Nor did I even consier the colour coding.
Well done.
I have wired split boards wrong myself when in compromising situations, read awkward, where the CU HAS to go where it was and you are on the top of the ladder in a hole with little light and p1553d off.
However, I have managed to find it in time, normally when one of the rcd's won't stay in!
Thanks Dave.
- Joined
- Nov 28, 2009
- Messages
- 14,762
- Reaction score
- 951
No, I only said the supply in to the CU was correct, not what happened to it once it got there. Didn't want to give the game away too soon. Sorry if that confused you
- Joined
- Nov 28, 2009
- Messages
- 14,762
- Reaction score
- 951
Actually, because my head torch was on, I hardly noticed it. I wondered why it got a little brighter, so I eased off the cutters and it went darker. I had to try it a couple of times before I finally realised it was the lead lamp coming on. Only then did I think WTF is happening.
Manator
©Honorary Essex Boy™
I understood what prodave meant, the cables had been terminated correctly, its just that the RCD had not been wired correctly, he indicated this in the early posts when he said getting warm, indicating a wiring fault, but which fault? When you are trying to work out example faults its always better to scribble circuits if you work back from prodaves initial experiance with the cutters acting as the connection, and the fact that a plug socket WAS being used (prodaves lamp), it makes working out what had gone wrong easier.
The biggest lesson this has shown is the absolute importance of the safe isolation procedure.
The biggest lesson this has shown is the absolute importance of the safe isolation procedure.
We know, can I have my tenner back?
remember, on a PME/TNCS system, you can generally regard the earth as being the same as the neutral conductor,
take into account also, this is not a 17th edition install, the lighting circuit is NOT RCD protected, therefore there is no 'imbalance' seen across L-N-E is this situation.
I hope no one objects to this comment as a pointer.
FFS, I missed a load of posts there. :|
anyhow, I thought the tails had been stuck into the outgoing(bottom) side of the main switch, with the neutral bars correctly in the incoming(top), so in effect, the same issue, just looked at from a different perspective,
I did see this once, I suppose the installer thought why bother to take them up behind the switch when the terminals at the bottom will do.
Rob_the_rich
Sock Muncher
- Joined
- Mar 2, 2010
- Messages
- 1,343
- Reaction score
- 27
Late night.. but if Prodaves bulb had blown while cutting he would have got >100A current with no RCD protection for a bit
no, he wouldnt. he said that was the only thing connected. if it had blown, then it would be open circuit, and nothing at the light cable would have been live. (unless there was some neons or some other loads connected that he didnt notice when he said nothing else connected)
Rob_the_rich
Sock Muncher
- Joined
- Mar 2, 2010
- Messages
- 1,343
- Reaction score
- 27
Sorry Andy, I meant that the high currents occurring whilst the bulb was blowing. Enough to trip an MCB or blow a triac, especially if the bulb was unfused. Unlikely event admittedly but would have made a molten mess of his side cutters. And the RCD wouldnt have tripped as the loads were balanced.
Manator
©Honorary Essex Boy™
It was a 16th edition board, main switch and RCD.
Rob_the_rich
Sock Muncher
- Joined
- Mar 2, 2010
- Messages
- 1,343
- Reaction score
- 27
oops checked diagram again, the lead light has a plugtop fuse.
- Joined
- Nov 28, 2009
- Messages
- 14,762
- Reaction score
- 951
The lead light has a plugtop fuse, but also, it's current path passes through the 32A MCB feeding the ring final it was plugged into. So no chance of 100A flowing anywhere. The circuit was as well protected from overload as it would be when turned on. (see diagram in "answer" post)
However, I very very much doubt the RCD would offer any protection. To work, they need a L and N supply to fire the trip coil (and the sensing electronics) But as there was no neutral, then If I had touched the L, or the floating neutral, I can't see that the RCD would trip. So in that respect it was less safe than if the circuit was properly energised. Also, if I had touched the floating N on the lighting circuit, I would have got a shock, but looking at the current flow, the RCD, even if it was working, would sense a balanced current flow, so would not trip.
It certainly taught me a lesson, not to rely on an isolator to isolate without actually testing for dead, and just because all the circuits appeared to have gone off, does not mean they have.
However, I very very much doubt the RCD would offer any protection. To work, they need a L and N supply to fire the trip coil (and the sensing electronics) But as there was no neutral, then If I had touched the L, or the floating neutral, I can't see that the RCD would trip. So in that respect it was less safe than if the circuit was properly energised. Also, if I had touched the floating N on the lighting circuit, I would have got a shock, but looking at the current flow, the RCD, even if it was working, would sense a balanced current flow, so would not trip.
It certainly taught me a lesson, not to rely on an isolator to isolate without actually testing for dead, and just because all the circuits appeared to have gone off, does not mean they have.
