Neutral-Earth fault defying logic...

[Spirogyra]

I hope it is only MY logic that is failing... ...and that YOU can figure it out!

This happens to be in Cyprus, but it could just as well be the UK or elsewhere.

Immersion heater (IH) connected to Neutral-Earth fault.

I installed a new one (Backer 314C, 3kW), ran for an hour then RCD tripped and would not re-engage. The previous one ran for 2 weeks then the RCD became un-engageable.
The RCD is a 300mA one protecting the whole property (detached villa).

When switched on, the RCD trips (usually immediately, but sometimes taking a few seconds).
If I bypass the original wiring and connect the IH to a 13A socket running off a different circuit from a different subDB, protected by a 30mA RCBO, that RCBO trips.
That happens even if the live is disconnected at the IH, and all 3 phases are disconnected at their isolators (there is a main DB and many subDBs), and all neutrals in the main DB are disconnected except the supply and the one serving the affected DB. That is: there is no load anywhere in the property (hence no means therein of raising the potential of the neutral), and the only connections running from the meter box to the 30mA RCBO are neutral and earth.
(the MCBs alongside the 30mA RCBO are all off)


Further info:

The system is TT or TN-S or TN-C-S (uncertain because there is an earth rod in the ground in the meter box, and the SWA supply is 4-core (no CPC core) and I don't know how the armour is connected upstream). Other villas in the development have similar supplies but don't have this N-E issue.

There is 1.6V across N-E at the IH (less in the meter box), presumably arising from loads upstream in other villas. For 1.6V to trip a 30mA RCBO, a resistance/impedance of 53 Ohm or less is needed. I can test the resistance of the IH N-E only with a basic multimeter, not a Megger; I get open circuit, ie. inconclusive. But this IH is only a day old. It seems soo unlikely that it would develop a N-E fault so quickly.
(other IH resistances: L-E infinite, L-N 19 Ohm (correct))


Last three thoughts:

In writing this, I realize I have not disconnected neutrals in the nearest DB. Although all phases are isolated (and there is no micro-generation on-site), I must disconnect them to ensure a low N-E resistance (<~50 Ohm) is not lurking there.

Other folks before me have been on-site and not found an explanation. They always end-up replacing the IH, and always the fault recurs soon. There have been 7 IHs in the last 3 months, no doubt bought from different stores.
I am loathe to replace it yet again without understanding the real problem!

If pushed, I would say:
- the cabling is OK
- something (specific to this one villa) is causing the IHs to fail prematurely. The only anomaly I have seen is mains voltage sometimes being 245V (the IHs are rated for 200-240V). Most of the times that I have measured, the voltage across the supply to the IH terminals has been 237V (off-load) and 227V (on-load), ie. within spec.

Any clues?

Thanks very much, for reading my post at the very least.

 

[Mikefromlondon]

1.if you just disconnect N as well as L at the IH, so that your IE is not connected to anything apart from Earth, what reading do you get between E & N, you said 1.6v and was that with IE connected to E & N or without N connected to it?

2. There can be a possibility that although you have disconnected all other circuits from the DB, but the power feed to your immersion heater circuit may also be feeding some other loads in the villa that you may not be aware of and any item on that circuit may be contributing to higher potential on E & N due to a leakage or some semi-fault condition whereby it is not enough to trip RCD on its own, and when you connect your IH you are also getting some leakage through it so that when both are combined it creates higher potential to exceed tripping current.
Some appliances use suppression capacitors between earth and neutral and earth and live that can cause leakage or unbalanced leakage when one capacitor goes leaky)

(one can check if other circuits may have been added to IH power circuit as you could could monitor any current being drawn when the IH is off or disconnected)

3. At what resistance setting on your multimeter did you check any leakage (resistance reading between IH L or N against earth) I mean did you try on your highest setting i.e on 200M, but it is likely to pick up noise on long wiring runs. That is if you did not have a meggar.

4. Is all the wiring dry and in good condition? as any leakages within the cable can lead you to think and everyone else that it may be the element. You really need to physically check the cable condition all the way from IH to DB.

5. If you are not able to physically check every bit of wiring as it may be bedded inside plaster or trunking or in conduit, then you could run a temporary wire run from DB to IH only via the RCD, just to rule out cable problem, for this you know what to do, buy a suitable size T&E 4mm I guess, and run it on surface just for the test duration, no people walking over it, and once you find that it does not trip when you do this, then it is likely the fault could be the cable itself or it may also be supplying possibly some other appliance along the path to this IH.

6. There could also be a problem with the RCD so don't forget to rule it as playing up and tripping for no reason, or tripping because of something else and you keep suspecting IH.

7. Have you measured leakages in the old IH that started to trip after 2 weeks, and was replaced, so that would be the one you should have checked for any leakages when it would not allow you to reset your RCD.

 

[Spirogyra]

Many thanks Mike,
I'm off out for the day, but near this particular villa, so I shall check issues from your post and update this later.
Re our 'temporary wire': in my 'If I bypass' sentence, you'll see I have already eliminated the original wiring and RCD.

 

[bernardgreen]

Neutral_Earth_trip

• bernardgreen
• 16 Jan 2011
• 48

 

[Spirogyra]

Bernard,
Thanks; good diagram and annotations.
Doesn't apply here, though, because there are no energized loads anywhere in the premises. All 3 phases are isolated at a bunch of 3-pole isolators (one per subDB) in the main DB. The only voltage is the 1.6-ish Volts on the neutral.

 

[Mikefromlondon]

Bernard's diagram is a good demonstration of the principle of RCD tripping during an E & N fault, but this is only good if the load it switched on, and drawing current, where the return current then splits between Earth and Neutral during E - N fault (shorting ) However, if you had other loads running from that RCD protected power, and if you were looking at your IH, which had been isolated on the live only (i.e. removing a fuse from the live feed but leaving neutral still connected) (both poles not isolated) and if you had a Neutral still connected, and there was an earth fault on that IH element, it would then trip the RCD because the return current from other loads (appliances) in use would be lower as some of it will pass through this IH fault directly via the earth thus bypassing some of the current returning through the RCD's Neutral sensor coil.


Spirogyra, yes you have a strange issue, have you tried different RCD? it could be faulty and tripping at less than designed current (30mA)

If so far you or others have changed so many elements, it may be worth replacing the RCD, probably be cheaper in the long run.

An RCD, it is basically a bit like a latching relay, in that it is latched ON manually by a leaver, and held on by a weak magnetic force that can be opposed by an electromagnet, and this electromagnet is composed of two opposing coils to generate a force to de-latch contacts held closed, however when current flow through the two opposing coils is equal, there is no net force as the force results in a zero opposing force to the latching magnet. But on a fault condition one coil in the neutral path starts to draw less current and the other in live path remains the same, thus creating an imbalance which yield a net force to oppose the latching magnetic force, and the contacts drop out.

So when you have no loads connected and only the IH, and it still trips , despite a new element, this indicates a possible fault with the RCD.

Other than this, if no other loads are connected, then Earth - Neutral fault currents can
be caused by induced voltages when lots of cables are running close to other heavy current carrying cables, but still as you said you had all 3 phases disconnected and it is still tripping, indicates RCD problem.

are there any power factor correction capacitors fitted to DB? could these be introducing induced currents by any chance? ( i am not sure if they can)

You have to ask, why are there currents flowing between earth & neutral when no loads are connected to Live? but that is only if your new RCD also starts to trip, and as you said there is no micro-generation like solar panels, or diesel generators or battery inverters that could cause this , so very strange, my feelings are that you need to try a new RCD.

 

[Mikefromlondon]

On another account, do you think that since even when you tried that new element (IH) on a different DB, with 30mA RCBO, and you said it trips just as well even when you have no live to it, but when you use that different DB, you should have no other load running through that DB as otherwise the RCBO is going to trip if the element is faulty, so now let us look at why does the element goes faulty after only a short run!

So I am beginning to suspect, that the culprit may be the element itself, perhaps manufacturing defect, so it effects that same batch you keep buying, have you also tried other brands, shame you can't really test any of the old ones probably discarded away, is there any possibility that they may be getting starved of water, so overheating and causing the OH cut out that somehow not only disconnects the live feed to the element but also then earths something and that in turn trips the RCBO?

or overheating causing damage internally and weakening insulation inside and that in turn causes fault currents and trips your RCBO, so I think you also need to monitor your water flow and filling system, to make sure there is adequate water flow into the tank as well as what is being drained out.

and water quality, can also be important, what if it was very acidic, and it attacked element in some way allowing water to ingress? all these possibilities should not be ruled out.

 

[Spirogyra]

Mike,
Thanks again.
I know threads such as mine take a lot of reading and concentration to get to grips with. I am always grateful to those who make the effort, even if they decide they can't contribute.

Unfortunately, I was not able to gain access to the villa today; will try again tomorrow.

The RCBO serves a double socket, into which I have connected my temporary connection to the IH.
I will check in the DB, but I am pretty sure it is just a single circuit (probably radial) coming out of the RCBO. But I will remove the faceplate of the double socket to check the circuit ends there.

No - there are no PF correction capacitors as such. ...Though there are motors which include capacitors (aircons, water pressurizing pump, 2 circulating pumps, pool pump, lift). All those were, however, disconnected on phase and neutral (when I had the subDBs' isolators off and neutrals disconnected).

Oh, and there have been no recent lightning strikes, so no obvious reason for (GPR) ground potential rise.

My existing testing has already used two RCDs - the 30mA RCBO in the boiler room, with the temp. connection, and the 300mA RCD in the meter cabinet, serving the whole property, which trips when the IH is not on the temporary connection. OK, there's a chance both devices are faulty, but one would think it unlikely.
I should add that I have been unable to get the main RCD to trip whatever loads (other than the IH) I switch on.
I have tried other loads in the socket served by the RCBO, up to 3kW resistive (ie. same current and same elevation to voltage on neutral as results from the IH); they cause no tripping.

Re some of your earlier questions:
- my meter is autoranging
- no reason to suspect cabling is not dry, but most of it is in conduit within concrete walls and floors so difficult to inspect. However, I have found that conduit runs can hold water for years, perhaps since open to the elements during construction...

I still have to:
- some measurements as you request, Mike
- disconnect neutrals in closest subDB (though this affects mainly the temp wiring)
- check no other devices on circuit served by the 30mA RCBO (ditto)
- try a new (3rd) RCBO/RCD
- perhaps find a way to monitor voltages (re reason for IHs going faulty):
- N-E voltage
- mains supply at meter (in case spikes causing IHs to go faulty, but doesn't sound right for various reasons)
- find the explanation, for the enlightenment of us all!

Mike

 

[Spirogyra]

Mike,
Re your second post today...

Bad batch issue...
I have just two IHs: the current one and the previous one. Both are the same make and model, but I don't know if they came from the same store. But I bought one from the same store recently and that has been in service for a while now, without problems. I can contact Backer, however, to see if they have had any reports of faults.

Water regime and overheating issues...
I think it is OK. The pressure relief valve (6bar) is in very good condition and has not been relieving. The automatic air vent passes air OK (tested when refilling tank). The thermostats on current and previous IHs look fine (OK, a bit of arcing evidence at the contacts) and the bimetal clicks OK. Temp set at 70C on both. The water is very hard; IHs here don't last long due to scaling and corrosion; 6-12 months is a common lifespan, but not days. The hot tank is fed by gravity from the cold tank on the roof 15m above, assisted by a pressure pump (and pressure switch) which work OK. There have been no reports of scalding hot water/vapour from the taps. The IH I replaced recently was scaly but not wrecked. So no reason to think the element may be running dry.

***Moderators note***

Some posts following this one have been removed as the advice contained within them was downright dangerous and I am disgusted that anyone would post such a thing on a public forum

 

[Mikefromlondon]

Do You mean that "Bucket test" advice, WITH CAUTION indicated was dangerous? Do you not think that when we are trying to fault find on electrical wiring or gear with live conductors and fault conditions that Danger is not present?

And it was not intended for brainless idiots AND only for the OP, who otherwise holds a good deal of knowledge and knows the dangers of working on Live gear, all electricians often have to work on live gear otherwise how are they going to find a problem which may only occur when live?

I am more disgusted with the attitude and approach that you have obliterated that thread rather than asking me to amend or edit some details.

 

[bernardgreen]

I have to agree that the bucket test is sometimes necessary to test immersion heaters.

but when it has to be used then the person doing the test has to know how to perform the tests and how to ensure safety.

The bucket test is not much different to boiling water in a plastic kettle, probably a bit safer as the water in teh bucket doesn't get to boil so scalding cannot happen.

 

[Mikefromlondon]

In future any sensitive advise will be PM'd rather than pasted on public forums, I was trying to help OP get to the bottom of his mysterious problem where he has had his RCD and RCBOs tripping one after another, his wiring tests have indicated no faults and I know a thing or two about elements having problems knocking RCBOs and RCDs even when new, Google it and you will see many people say they bought an new oven and it keeps tripping them off, so what do you do?

Run them in an UNSAFE mode by not wiring them through an RCD, or use an isolating 5Kw transformer, or simple remedy run them for a while without an RCD until element purges out any moisture. Which is what I did and ever since its been running happily not tripping my RCD.

 

[Spirogyra]

Thank you, Mike and Bernard, for your continuing posts.
I am sorry some upset has occurred. I dare not enter into it, especially as I am new to this forum.

Back to the technicals...
The site manager and I got electrical contractors in (the ones that did the original installation), seeing as they know the wiring best (one hopes), they have access to the fullest range of testing tools, and will have responsibility for any problems with the wiring if due to installation flaws. The particular electrician was sure the problem related to using a 14" IH rather than an 11inch IH. He replaced the IH (the one I installed recently) using an 11inch one (still s.steel) by the manufacturer of the original (a few years back) - Heatrae, who also supplied the hot tank. I told him 14" was not an issue, as there was about 3" gap between the end and the side of the tank (the IH sits horizontally in a vertical tank). While he was installing it, I surveyed the other villas: all of them have 14" models, including the Heatrae ones (which were copper not s.steel, but that's another issue).
So, I have handed over responsibility to him, but the site manager will keep me updated. As & when it is fixed and we have an explanation (hopefully!), I will post here, as I know you good folks are interested.

 

[bernardgreen]

There is 1.6V across N-E at the IH (less in the meter box), presumably arising from loads upstream in other villas.

This may be the mechanism, a fault Neutral to pipe work or other metal work that has a connection to the ground

 

[Mikefromlondon]

Bernard, yes that is strong possibility he should check what type of earthing system there is apart from a ground rod that he said was there, really needs to check if the armoured cable earth is tied to N at the DB as well or not.