Electrics: Earth kick

[garryb59]

I think I'm finally at the point of asking me mate Dave [sparks] from the pub to give the scene a proper once over, but just before I do, I'll like to ask anyway.

So...I installed a new[Crabtree unit off Ebay] split load CU in my parents Park Home at the beginning of the year. Non protected side = 2 x 6amp light circuits, 1 x 32a circuit for boiler and fridge freezer, protected side [30ma]: 2 x 32a socket circuits and 1 x 32a oven circuit. This is a small residential Park Home.

Earth is TN-C-S

Problem: Nuisance RCD random tripping. Remove all appliances, check sockets, check wiring etc....nothing obviously visible. RCD still trips, sometimes from pc, sometimes tv, sometimes kettle...but nothing obvious...etc

Temporary solution: put offending circuits over to the non-protected side, but leave Oven circuit on RCD side. Oven has never tripped RCD since.

Having left it like this for too long: Father does some tiling in the kitchen and gets a considerable shock from the earth on one of the sockets. This particular socket is on the ring, but has a spur feeding a washing machine. Jumping to uneducated solutions we guess it's the washing machine [which is new] that has been causing all the problem. Service engineer comes and passes the washing machine as being fine, but says he can detect a voltage in the earth with his meter [I wasn't there at the time].

I have had all the circuits out from the CU, and done simple continuity checks, I've checked sockets, wiring where I can, earth connections from the meter back to the CU, but can find nothing. I'm clean out of ideas.

I suspected the RCD might be faulty originally, but the fact that it hasn't tripped at all with only the oven, and the fact that my father got a kick from the earth points to a different kind of fault.

Like I said, I think I'm almost at the point of handing it over to higher powers, but is there anything I might want to look at?

Appreciate any input.

Cheers Garry

[jim_in_sussex]

unlikely with a good brand, though not completely impossible. much more liley you have an earth leakage fault. Possibly the fault was there long before you upgraded to the new CU & it is just that the new CU is doing its job & shows up the problem.

but the fact that it

In simple terms RCDs trip because the supply current flowing thru L (line) is not equal to the return current flowing back thru N (neutral). The assumption is that discrepancy is due to a diversion (leakage) of current via earth back to the supply source. Common values to make the RCD trip are earth leakage currents of 10, 30 and

100mA.

Many appliances - especially electronics- have a natural earth leakage current of a few mA but in normal circumstances the total is not enough to make an RCD trip. What takes it over the top can be something like a nailed cable fastener into a piece of wood [a reasonable but not perfect insulator] just grazing the N conductor.. Whether it goes over the top (ie trips) at any particular time when there is an installation fault to earth is thus dependant on the varying level of natural earth leakage from correctly plugged in appliances. This is the basis of nuisance tripping.

Except for gross faults RCDs really continuously monitor the general quality of the earth separation from L & N.

In contrast MCBs trip [and fuses blow] due to current overload - either due to a sudden short cicuit or due to too many or too power hungry appliances being used at the same time.

The most likely faults to look for from your description of the problem are a break in earth continuity; a measurable resistance from N to E or on L to E - perhaps indicating an insulation fault.

As there is enough voltage to give a shock that might indicate an L - E fault. Possibly it could be more complex, eg an L-E fault plus an earth continuity problem.

Start by disconnecting all cables at the CU [after disconnecting all appliances] and test for mutual infinite resistance between them all.

HTH .

[Andy Wade]

A "voltage in the earth" is rather an unhelpful description. A voltage has to be /between/ two points - which two points are they?

A shock from the earth of a socket where a washing machine is involved immediately suggests the possibility of a open circuit (broken) circuit earth connection somewhere. The 'floating' earth would then take up a voltage of about 115 V due to the capacitors in the EMI filter in the machine, and would be able to deliver enough current to be felt as a significant shock.

I suggest you get a long test lead (a piece of 3-core flex will do, with all cores in parallel) and connect one end to the main earth terminal or the earth bar in the CU. With the relevant circuit powered down and all appliances unplugged, work round all points in the circuit with your low-ohms meter (or battery and bulb) checking for good continuity between the 2nd end of the test lead and the earth contact of /every/ socket on the circuit.

[Christian McArdle]

You need to do a full circuit test with the proper tester. It must support continuity measuring with 200mA test current. It must support insulation resistance testing with 500VDC. If you don't have such a device, then I'd suggest getting an electrician in to do a full installation test, as proper test equipment such as this (and the RCD tester you will also need) cost hundreds of pounds. This equipment will find faults where a standard electronic multimeter wouldn't have a chance.

Christian.

[meow2222]

I would let a qualified spark deal with it, as you dont seem to have the knowledge needed to get the knowledge to fix it, and this is one type of fault that can certainly kill. I sure hope you dont have an electric shower.

They would do continuity tests on the earthing, use a megger to find the earth leak fault(s), and hopefully recommend a better grouping of circuits on each side of the CU.

Meanwhile you could test all your appliances with a multimeter to look for leakage between L/N and E.

I would remove power from any dangerous circuit until its fixed. A ring that has a live earth conductor is a serious hazard.

NT

[Dave]

This was almost certainly the effect of normal earth leakage currents from something that was designed to be earthed - BUT WHICH ISN'T. If the shock really was from an outlet (rather than a piece of equipment) you have an earth fault in your building wiring which NEEDS TO BE FIXED PDQ. You should really isolate the ring until you can fix it.

Dave

[Bob Watkinson]

In addition to the tests already described you need an Earth Loop Impedance test. When you check continuity you measure the earth path back to your CU. An ELI tests the whole earth path back to the suppliers transformer. This is vital to amongst other things to ensure your main earth terminal is actually connected to earth.

[garryb59]

I have a basic DMM from Maplins [which I know is insufficient for anything other than very basic tests], and what happens is: if you connect the MM between the earth of the washing machine spur and either of the other two ring earths you get about approx 50v reading, if you then plug the washing machine into the socket, the reading goes up to 110v.

Would this tell me anything more than testing for continuity between the two disconnected earth wires at the CU? I ran this basic test across all pairs of L N & E wires at the CU of both rings and all appeared to be unbroken - if a noise signal on my MM is valid?

Garry

[garryb59]

Thanks for all these replies, appreciate the time.

I don't have the necessary eqiupment to carry out the proper tests, so I've come as far as I can I my own. Time to get help.

I'm not looking to bodge anything, just a curiosity sucker looking for a bit of understanding that's all.

cheers Garry

[Dave Plowman (News)]

Sounds like the earth in the spur is broken. Can you measure it from the socket where it's spurred from? If there's volts between them turn off at the mains and measure the resistance. It should be near zero ohms for a short run.

The chances are the earth's broken inside either the socket it's spurred from or the spur socket itself.

[garryb59]

Hi Dave,

After the washing machine guy had been, we thought it might be the cable, so we replaced that [about 800mm length], but still the problem remains.

Garry

[dave]

Am not trying to be a smartass, and what you did is very tempting - but it could have killed your father! Not worth it really is it. As the others have said it does look a broken earth to the spur.

[Dave Plowman (News)]

Have you physically checked the earths inside the spur socket and the one it's spurred off on the ring?

The washing machine has likely mains interference suppression filters which go to earth so it's no surprise if the volts come up on a broken earth when it's plugged in.

[garryb59]

Of course, just as he would have been killed with the old fashioned non-rcd CU's, good job nobody's got them anymore. God forbid.

Which has been changed twice. Not so.

Anyway. Help from a better mind is at hand...

Garry

[Blueyonder]

Your father received a shock. He could have been dead now. Why delay having another pair of eyes looking at this even one more day?

[Andy Wade]

Yes, that's exactly what you'd expect if you break the earth and insert the meter. The 50 V is just capacitive coupling between the floating earth wire in the spur cable and the nearby 'live' wire. The 110 V when the washing m/c is plugged-in is due to filter capacitors in the machine establishing a sort of centre-tap between L and N (IYSWIM).

Well I suggested it as a simple way of checking that every switch, socket, FCU or whatever is actually earthed - "continuity of circuit protective conductors" in the jargon. A continuity test between two ends of a ring at the CU only tells you that there is a ring /somewhere/. When spurs and possibly various kinds of 'illicit' extensions to a ring are involved, it may well be possible to measure continuity at the CU ends but still have parts of the circuit unearthed, so to speak.

[garryb59]

Ok.

Andy, I did what you suggested, took a long lead connected to the earth block on the CU and got a reading from earth on all the sockets and switches.

Anyway, I've got some help on the way now, so I'll see what comes to light, and I'll let you know. Thanks for the advice offered here.

Garry