An electrical puzzle.

I wired in a new cooker hood today. The old one was wired to the mains via one of those fused wall switches with neon indicator. I prised the switch cover open enough to replace the old wires with the new and left it like that while I turned the power back on at the CU. So far so good, but when I (electrically) closed the wall switch there was a bang and some smoke, followed, maybe half a second later, by another bang and a bit more smoke  from the wall switch.. It turned out that the insulation on the wires to the neon indicator had failed resulting in a short, presumably when I prised the cover off.

When I went back to the CU both the MCB for the circuit and the RCD for that half of the CU had tripped.

My problem is that there were two bangs separated by far more than the reaction times of either the MCB or RCD. One or other should have shut off the circuit after the first bang surely? Any hypotheses about the series of events would be interesting!

Robin.

Edited By Robin Graham on 08/04/2023 00:49:01

The RCD does not provide phase fault protection such as short circuits. The original fault was a line/neutral short, tripping the single pole MCB disconnecting the live only. Subsequently the neutral conductor may have touched an earthed part and would then trip the RCD. The neutral voltage to earth is generally sufficient to drive the RCD with more than 30 mA and hence tripping the device.

I have tripped an RCD on a circuit that has been isolated at the MCB by cutting through a wire and shorting the neutral to the earth conductor. As Willem says there is enough voltage in the neutral wire to drive the RCD into tripping. The neutral return has a small but non-zero resistance. Current still flowing through this resistance follows Ohm's Law to produce this voltage relative to earth on the neutral wire. The original live to neutral short probably produced enough heat to remove more of the insulation that had degraded. This allowed the second fault of neutral to earth shorting.

Martin C

I would think the first ‘bang’ was the neon connection fusing - but not taking the 45A (?) breaker - and the second was line to earth connection, possibly due to a reconnection due to an already melted conductor making a further short circuit? Just guessing, of course.

Edited By not done it yet on 08/04/2023 08:39:50

There are usually a couple of volts on the neutral conductor in house wiring, and this can be enough to trip an RCD if the neutral is touched to the 'earth' conductor.

The first time it happens when you are working on an isolated circuit; and trip the RCD, plunging the house into darkness, you think "I am sure I switched the power off to this circuit, and checked". So you triple check, and yes, you did. But those few volts on the neutral can be enough to trip an RCD.

PS: ALWAYS double check that a circuit is safe before working on it. Simply switching off the MCB is not enough - you need a second independent check that all conductors of the circuit are safe, e.g. with a decent multimeter.

I fully endorse checking properly for isolation but in an ideal world a multimeter cannot be recommended due to the increased chance of user error. Back in the real world it is unlikely that any of us will have a tester and proving unit so we are back to the multimeter that many of us are likely to have. Check that the multimeter is set correctly by testing a known source then check L-N, L-E, and N-E then prove the multimeter is still working on a known source, of course if you have turned all the power off rather than one circuit the source may now be off which is why a proper tester has an independently powered proving unit. Although the proper test gear which is expensive and unlikely to be in anyone’s toolkit who isn’t a working electrician the process of checking is valid with whatever test equipment you have. I find it difficult to recommend shorting cables to earth to ensure they are isolated as if you get it wrong and it is still alive the flash and bang and the possible expulsion of molten copper could injure you. This may be marginally be better than electrocution. If you totally lack any test equipment then working on mains voltage equipment should be avoided.

Mike

Hi As the earth is bonded to the neutral ,there should not be any voltage between them.

Maurice

"Should not" agreed, but there often is. I have measured 19 volts on a system which had passed an earth loop resistance test.

George

Even so, as Gerorgineer and I both say; there might well be volts on your neutral. Have a look if you don't believe me ! Quite possible for a few volts to be present on a neutral wire, trust me, I have measured it many times.

Re Multimeters, yes, well I suppose I was assuming reasonably competent operators. Having been trained and worked with electronics all my life, I check and double check such things automatically.

I certainly wouldn't short any cables to earth as a first check though !! And I try not to short anything to anything, for the reasons stated earlier about tripping RCDs.

I'd go along with that. Fuses (including the conductors of a neon lamp) tend to go, "bang", but RCDs, etc. go, "click".

Not really, "bonded", and not always in any event.

A PME or TN-C-S supply has a combined neutral and earth supply, which splits into N and E at the service head and the nearer you are to the origin, the lower the voltage between them.

In a TN-S supply, N and E split at the transformer. In a TT supply, E finds its way back to the transformer via the general mass of the Earth and a rod at either end.

Then under certain conditions, you can have diverted neutral currents, but that seems unlikely in this event.

All a matter of scale. The earth leakage trip may be 30mA if the wiring is 0.1 ohms it only takes 0.003V to generate 30mA of current. Triple that for N-E loop resistance of 0.3 ( earth conductor is smaller)ohms and you get 0.009V. A load of 0.1A will generate 0.01V in a 0.1 ohm neutral.

So if there is pretty much any load on the ring connecting neutral to earth will trip the leakage device.

Robert.

Yes, he said both bangs were from the switch - a distace from the RCD and/or circuit breaker in the consumer panel.

Try working out the current using Ohm’s Law when the resistance is zero. Voltage will be zero, but the current will be indeterminate.

Edited By not done it yet on 08/04/2023 20:14:32

Well, a PFC of say 1000 A would make any fuse go, "bang".

I might add that unless you are in a heavy industrial unit with its own transformer, the resistance will be measurably greater than zero Ohms.

Thanks. I think Willem explained it in the first reply. Annoyingly I was aware that a neutral / earth short can trip the RCD having had exactly the experience which John Doe 2 described, but didn't think of it before posting. Perhaps surprising that that there was enough neutral/earth current to cause the second bang, but certainly not impossible.

Doing this I had a minor brainwave when faced with the problem of threading the new flex behind the kitchen units. "Ha!" I said to myself, "join it to the old flex with a piece of that adhesive heat shrink tube and pull rather than trying to poke it through."  When my wife saw me making the join she asked " but how can you be sure that the red lines up with the red and so on?" I was tempted to mystify, but came out with the boring truth. Anyway, it worked a treat.

Robin.

Edited By Robin Graham on 08/04/2023 23:36:56

I think there are too many people on here trying to (wrongly) explain how an RCD works. it is nothing to do with neutral to earth voltages .To put it simply an RCD compares the current flowing in the live conductor with the current flowing in the neutral conductor. If they are not the same then there is current flowing through another route. This route is to earth. Household RCD devices are set to trip when the current comparison exceeds 30 mA of a difference.

A single RCD in a distribution board usually protects a number of circuits in the house/workshop so, because the neutral wiring is all joined together at the distribution board then earthing any one of the neutral wires anywhere in the house (even on a circuit that has its live feed isolated) will cause current to flow to earth from any of the other live circuits, thus causing the RCD to trip.

Martin

Good point, but a few volts present on a neutral - even in a circuit isolated from the line supply - when shorted to earth can cause enough current to flow to make an RCD see a sufficient current imbalance to trip out.

The potential difference, (in volts), between a neutral and an earth is what makes the current flow !

Martin C is correct in saying that the RCD measures the difference in flow between the line and neutral conductors, but in this instance, there is no current in the line conductor because its MCB has tripped.

As John Doe says, there must then be sufficient voltage between the neutral conductor and circuit protective conductor for a current to flow.

An RCD with a nominal current of 30 mA does not magically trip at exactly 30 mA. It must do so within 300 mS at that value, but will trip faster at a higher current and should not trip at 15 mA or less.

Edited By Chris Pearson 1 on 09/04/2023 10:32:35

Edited By Chris Pearson 1 on 09/04/2023 10:33:04

Interesting to hypothesise about where volts between neutral and earth come from. I offer this example:

In the circuit, to keep the maths simple, a 250Vac source is feeding a 100 Amp load through a cable with a 0.1 ohm resistance. If a voltmeter is read near the Neutral/Earth connection at position V1, the voltage will be close to zero. However, if the voltmeter is moved to position V2 at the load, it detects the voltage drop caused by cable resistance. Though about 10V above ground, there's no current imbalance at the source, so an RCD is happy. So is an earth resistance check. This example is all good I think (other than the power wasted in the cables!)

Another possibility is that the multimeter is detecting RF on the line, perhaps from a Ethernet over Power Line system. These piggy-back on mains wiring to carry a modulated radio signal through buildings. Power equipment doesn't react to the signal, but it can be detected by suitable electronics. A multimeter might 'see' both signals. Low frequency mains (50Hz), pretty much obeys ohms law inside the meter, which measures it correctly. Radio signals don't obey ohms law in the same way: high frequency AC sees the meter circuit as a mix of resistance, inductance, and capacitance, which allow it to bypass the meter's dropper resistors. The movement of an analogue meter is protected by diodes, which rectify the signal and produce an uncalibrated meter reading. Diodes in a digital meter can also cause anomalous meter readings. This cause is also OK - only the meter is confused.

My circuit diagram deliberately shows 3 earth connections, which is not normal! I did it to stimulate debate about earth faults - what happens when the earth system malfunctions, either by developing resistances, leaks, or part of it goes open circuit. Various strange things can happen, and there's at least one supply-side earth fault where consumer Neutral becomes Live with spectacular results.

And then there are electrical storms and solar activity...

Dave

Edited By SillyOldDuffer on 09/04/2023 11:18:18