RCD on new consumer unit

[dmccormick]

Thanks to all for your advice. If you're interested, here is what I did in longwinded format; if not, the summary is that it is mostly fixed with one remaining doubt.

After reading a few faqs about wiring and CU's I felt confident enough to have a go. I turned off the double red switch and started by simply checking the RCD neutral bar against the earth bar and found its resistance was essentially zero. Repeated this for the non-RCD neutral which was also zero.

Then I realised that in order to bypass the RCD they had connected both neutral bars together and both buzz bars together. I removed these interconnects, put the safety cover back on and switched the power back on.

I set all MCB's on and then tried to set the RCD, but it immediately tripped (no surprises there). I then turned all the RCD MCB's off and tried them one at a time with the RCD on and found that only one of them caused it to trip. However, I didn't know which circuit it was. It had two wires going into it and either of them caused the trip and so I thought it was probably a ring main as it was on the RCD side and was not the really thick cable (which I guess goes to the cooker). With this MCB pulled still I checked the sockets downstairs and they had no power so I assumed this is where the problem lay.

I therefore turned the main switch off again and rechecked. The neutral/earth resistance: for the RCD side was still zero, for the non-RCD side it was 400 Ohms.

I removed each neutral wire in turn and checked it against the earth. On the RCD side I found 1 large wire (ie. mulitple strands) had a zero resistance and all the others were infinite(well off the scale). On the non-RCD side I found 1 wire with 400 Ohms and all the others were infinite.
I then started to check all the downstairs sockets resistance from the earth pin to the neutral.

At first I thought there were no problems - I was testing resistance from earth pin to neutral with my meter set to beep mode and it was dark. However, I then noticed some of the sockets were giving around 350 ohms resistance which was not enough to cause my meter to beep, but obviously is not right.

I went back to the CU and removed both the suspect RCD side neutral wire(0 ohms) and the non RCD side neutral wire(400 ohms) and went back to test the sockets. The 400 ohm sockets were now OK. Eventually I found several others with zero resistance.

On opening them I found 2 of them had unshielded earth cables that were actually touching the neutral cable exposed copper part. I put plastic earth sleeves on them and trimmed and reconnected the neutrals so that no bare copper was extending from the terminal.

Having done this and put the lot back together I was dismayed to find a N-E short again on one of them. Close inspection revealed a nick in the neutral cable that only chafed on the metal box when the socket was screwed in.

I replaced the damaged wire and put everything back together and it all now works - the RCD sets and the Test button makes it pop.

The only thing I am left wondering about is the 400 Ohm N-E fault on the non RCD side which I thought would be a lighting circuit. However, having written all this I now realise it must be connected to one of the downstairs rings where I saw the 350 ohm faults before. Then again, maybe I tested those sockets whilst the neutral bars were connected together with the zero ohm RCD neutral disconnected, in which case I guess all neutrals would have shown an earth fault.

More to investigate tomorrow.

However, it seems that what the guys who did the job an Saturday said was fair and true. Will probably get them to look at the remaining N-E fault and also to fix a few DIY shockers from the previous owner (outside socket wired up via a single pole light switch to an indoor socket, metal light switch inside that has no earth connection)

 

[ban-all-sheds]

Close inspection revealed a nick in the neutral cable that only chafed on the metal box when the socket was screwed in.

http://www.mccormick.plus.com/socket.jpg[/QUOTE]
What a truly disgusting sight that is - damaged insulation, unsleeved earths, no grommets. If this is typical of the rest of your installation expect more fun and games......

 

[Damocles]

Sounds like you are getting into this. Remember that a N-E fault on one RCD circuit can still cause RCD trips even when its mcb is switched off. In fact it can cause RCD trips even when all the RCD MCBs are off.

 

[dmccormick]

[ban all sheds wrote]reeeee[/ban all sheds wrote]

 

[dmccormick]

sorry for lame post just then - I was trying to include a quote from ban-all-sheds "is the rest of your wiring like this", but didn't know how to do it and pressed the wrong button by mistake !

Anyway, unfortunately yes the downstairs wiring is fairly typical like this with bare earths. I looked at the new stuff they put in upstairs and it all looks very neat with grommets, earth covers, etc.

I also located the 400 Ohm fault - it is on the immersion heater circuit between the switch and the heater. However, that cylinder and heater is being removed on Weds and the wiring will be dicsonnected as the new one is going in a different place and they've run new wiring to it.

When we can afford it, we'll have the downstairs rewired as well.

For now I just need to run an earth cable to the light switch downstairs that has no earth (especially nice as the switch has a metal face plate!). Is there any way to do this without having to break into the plaster ?

 

[ban-all-sheds]

I also located the 400 Ohm fault - it is on the immersion heater circuit between the switch and the heater.

Is it actually in the wiring, or have you got an earth fault via the element - it does happen.

When we can afford it, we'll have the downstairs rewired as well.

Try to afford it sooner rather than later....

For now I just need to run an earth cable to the light switch downstairs that has no earth (especially nice as the switch has a metal face plate!). Is there any way to do this without having to break into the plaster ?

It depends on how the cable was installed. If it's just buried in the plaster, then no way. Under capping you might be able to pull a new one through, but I wouldn't bet on it. In conduit, easy peasy. Chances of conduit having been used? Vanishingly small.

So does the lighting circuit have no earths, or is it just that some muppet has cut back the earth core at the switch?

 

[dmccormick]

Is it actually in the wiring, or have you got an earth fault via the element - it does happen.

Now that you raise the question, you've answered a niggle I had. The circuit in question has a double pole switch with red light downstairs, upstairs is a normal 13A socket with a plug in timer and then a normal 13A plug connected to the immersion heater. I disconnected the neutral and live at the CU and tested the N-E resistance in the switch. With it open there was a fault on the LOAD side only. I then went upstairs and checked the socket, but couldn't find anything wrong (this socket had previously almost blown up due to a shorting live-earth cable and had been cleaned up). Having reconnected everything I went back to the CU and checked the N-E resistance before reconnecting to the CU - again the N=E fault was there. Must be that the fault is as you say in the element or even the plug as these were of course disconnected when I checked the socket.

So does the lighting circuit have no earths, or is it just that some muppet has cut back the earth core at the switch?

Fortunately the latter - there is no earth cable in the switch in question. I checked(by putting my meter between the earth pin of a socket and the light or switch) the lights themselves and other switches and they are earthed.

 

[plugwash]

NE faults usually go unoticed on a non-rcd system (unlike live-earth faults which are going to blow the fuse and thereforge get fixed)

so when putting a rcd on an existing cuiruict (ie a CU change) this problem is pretty common

 

[dmccormick]

Thanks again to all for your advice. I have just one more question if I may:

I worked out the situation with the non-earthed light switch and think I know where I can connect the necessary earth. I want to run a good earth from a socket and break it into the earth wire in a 2core and earth flat cable. I was thinking of running a single green/yellow earth wire from the socket, across 2 joists to meet the 2core and earth wire, cutting the latter and then joining red/red black/black and the 3 earth cables in some kind of junction box. Question is whether it is OK to run a single earth/lead and what sort of junction box I need ?

For info, here is the story with the light switches:
Switch in lounge is 1way 1 gang, controlling 1 light. It has two cables running in to it, each of which contains only a single black wire and there are no earth leads - this was my original concern.

This switch is on a shared wall with the dining room(DR) and back to back with it is a 2 way 1 gang switch for the 2 lights in the dining room. The dining room switch does have an earth lead, so by running a wire straight through the wall linking the other switch I thought I had it solved. However, here is the nasty bit – although the DR switch has an earth lead, it isn’t earthed (tested by checking against a known good earth).

The other switch in the DR is 2 way 1 gang, also with earth lead that is not earthed. The earths are however linked – i.e. both switches share a common earth circuit that is not itself earthed.

What was even more puzzling is that all three lights themselves are earthed properly.

Here is a description of the cables in the switches and light fittings:

DR switch 1:
A 2c+e cable connected as: red-L1, black-L2, bad earth to plate; a single core cable connected black-C

DR switch 2:
A 2c+e cable connected as: red-L1, black-L2, bad earth to plate; 2 separate single core cables, both connected as black-C

Lounge switch:
2 separate single core cables, 1 connected to Live, other to ~

DR light 1 ceiling rose:
A 2c+e cable connected as: red-brown, black-blue, good earth to light; a separate single core cable connected as black-brown.

DR light 2 ceiling rose:
A 2c+e cable connected as: red-brown, black-blue, good earth to light; another 2c+e connected as black-blue, earth to light, red – bent back up into ceiling and connected via a plastic terminal connecter somewhere I can’t see !

Lounge light ceiling rose:
Two 2c+e cables connected as red-red(not to light at all), black-black-blue, good earths to light; a separate single core cable connected as black-brown

How I think it is wired is:
A 2c+e cable links the DR switches L1/L2/Earth, but the earth goes nowhere else.
A 2c+e cable links the DR lights
The C from DR switch 1 goes to the red wire linking the 2 lights.
The second 2c+e cable in DR light 2 rose comes from CU and gives neutral supply and earth to the lights, its live lead is split off and fed via a single cable to the C on DR switch 2.
The second single cable in DR switch 2 C terminal is fed to lounge light switch as live feed.
The other cable in lounge light switch goes to light (switched live)

I.e. a complete dog’s breakfast !

 

[plugwash]

are we talking double insulated single cables here (ie colored insulation and then a grey or white sheath) or are we talking cores with just a single layer of insulation (which shouldn't be run seperately without conduit in the first place)

its very bad practice to staeal an earth from another cuircuit in case that cuircuit is later removed but if you can't get a cable through to a light fitting that has an earth then its better than nothing

single core earth cables are allowed provided thier conductor size is at least 4mm and a junction box would be a way of breaking one off from a cuircuit

 

[dmccormick]

Thanks plugwash. Yes the single cables are black and then have a grey outer sheath. Are there different types of junction box or just one ?
I know it's not great to steal the earth of another, but when we have the downstairs rewired I'll point it out to the electricians.

Thanks for all your input.

 

[Damocles]

Thought the minimum size for a flying earth in bonding applications was 2.5mm provided it has mechanical protection, which I interpret as meaning so people cant fiddle with it and break it off. Otherwise 4mm. For internal earth bonding, as appropriate for the circuit.

 

[plugwash]

i interpreted mechanical protection as protection from damage by tools (like for burried cables) so i would have thought that for direct burial or running in the void of seperate earths 4mm would be needed

as for the junction box since you will be putting it in a ring a 3 terminal 30A is the type you want

 

[dingbat]

Ordinary twin and earth has 'mechanical protection' in the form of the grey PVC outer sheath. X-grade yellow/green is, of course, not thus protected, so in most cases you should use 4 sq mm or else bury it in plaster, or cap it or use trunking, conduit, etc... I've never seen green/yellow cored sheathed single cable but if it exists then you could use 2.5 sq mm in that!

 

[dmccormick]

All sorted now and thinks to everyone for their help.

Only 1 niggle - I ran the earth from one of the new sockets that the guys installed last weekend. I noted that their earth cables go into the socket plate, but there is no connection back to the metal box. Obviously the metal screws will form this link when the plate is attached, but is it a building regs or IEE requirement to have this earth link cable back to the metal box ?