Earth leakage problem. Help.

[Derlyn]

Ok. Never before had this problem.

Connect supply live and neutral to earth leakage.
Press test button and it trips. Cool.

Connect load live to bottom of unit. Press test button and it trips. Cool.

Connect load neutral to bottom of unit and press test button. Does not trip. Not so cool.

When I connect the load neutral to the unit, it does not trip.

Come now, my brothers, I need some more brains here.

[Derlyn]

Should have said, when I connect the load neutral AND THEN PRESS THE TEST BUTTON, it does not trip. It also does not trip with the plug tester. Tried 3 new earth leakages, same problem.

[GCE]

Take it that it is single phase unit and that voltages appear correct including neutral to earth been almost nothing.
Is your earthing test out OK on Earth Loop impedance

Definitely strange one - It should pick up the imbalance as all it does is check what goes down live comes back on the neutral in simple terms.
I seem to recall a similar problem on an ABB ELU a couple of years back and the niggle in the back of my head recalls an earthing problem, in that the return from the fault went back on the neutral not down to earth to create the imbalance - I need to ask the guys in the morning.

I would take amp reading on live and neutral and see if they are balanced and if you have milliamp scale see if the imbalance appears with ELU tester , next turn off all live circuit breakers and try the test button

Other plan would be to connect a constant load directly to the unit , 150w lamp or stove plate , and short neutral to earth or place a plug tester in parallel directly under the unit

[Justloadit]

The way that the ELU works, is that the Live and Neutral wire go through an iron core. The core then has a secondary winding connected to a trigger mechanism. When the amps going through the Live is returned through the Neutral, the current difference is zero. When there is an Earth leakage on either the live and Neutral, then the current in the Live and the Neutral is not the same and does not cancel out, and the difference then if greater than 20mA for a 20mA ELU, triggers the mechanism, and it trips.
This only works if the Earth connection to Neutral is done before the ELU and has good mechanical connection. However if the Earth to Neutral connection is lost/broken, then the ELU will not detect a leakage to Earth, as there is none being the missing link from Neutral to Earth. This is the same scenario when connecting a Genset to the ELU with out a Neutral/Earth bridge, and the reason that the SANAS states that alternative supplies require their own Earth/Neutral bond, else the ELU does not pick up an Earth leakage fault..

[Derlyn]

The test button on the earth leakage relay should work irrespective of whether the neutral is earthed or not.

There is an internal resister that is in series with the test button and when the button is pressed, a current will flow through the core to the resister but the return current bypasses the core to the neutral thus creating the imbalance that trips the unit. The test button operation thus does not rely on an earth for the unit to trip.

I'll be back at site tomorrow to attempt to find the fault. At the moment, the downstairs plugs do not have earth leakage protection, but the building is not yet occupied.

I need to let what little brains I have, rest for the night and persevere tomorrow.

I'm still puzzled. Usually one has nuicance tripping but in this case, the unit does not trip when it should.


This following diagram shows how the test button is internally connected in the earth leakage relay.
Now, something on the installation is preventing the test button from tripping the unit.
I have to find out what it is.

[Dave A]

Clutching at straws here, but -
Earth and neutral is bridged on the load side?
An earth neutral fault to the point where the neutral loop impedance on the load side is a tiny fraction of the neutral loop impedance on the supply side.

[Derlyn]

Good morning Dave.

On my way back to site now.
Some more testing to do.

Theoretically, a bridged neutral and earth on the load side should cause the ELR to trip if there's a load present, not prevent it from tripping, which is the case here.. That's the confusing part.

Hopefully, later today, I can give some feedback.

Like I said, it's a first for me, so it's one of those occurances which I am sure doesn't rear it's head too often.

[Dave A]

On reflection I am possibly wrong on the relative neutral loop impedance readings. But I am mindful of your typical earthing arrangement in East London and that the unit is vacant - i.e. possibly no load on the load side to produce a current to leak to earth on the load side.

[Derlyn]

Ok. Was a neutral earth fault on a 1mm FTE going to an outside bulkhead light.

The live was not yet connected in the db because the light switches were wired but not yet mounted due to the painters still doing their thing.

What is puzzling is that this fault caused the elr not to trip, instead of tripping when there's a fault.

Granted, there's no load being drawn there, but this is the first time ever that I've come across a neurral fault rendering the elr useless.

Maybe someone can explain this to me one day.

Lekker weekend all. ✌️

[GCE]

Was the light on ELU ?

[Derlyn]

@GCE. Yes.

[Derlyn]

YoU will not believe how many circuit sketches I've made to try and work out how that neutral earth fault can render the test button inoperable, without success. I once spoke about an earth leakage not tripping on an earth fault if there's no load on the output. I even made a little video to prove it. When I have some spare time on my hands, I'm gonna play around on the workbench and see if I can replicate what has happened the last 2 days. Theoretically it shouldn't be possible, but believe me when I say that is what happened. I , however, cannot explain why.

[Dave A]

I have to find out what it is.

[ATTACH=CONFIG]8907[/ATTACH]

YoU will not believe how many circuit sketches I've made to try and work out how that neutral earth fault can render the test button inoperable, without success. I once spoke about an earth leakage not tripping on an earth fault if there's no load on the output. I even made a little video to prove it. When I have some spare time on my hands, I'm gonna play around on the workbench and see if I can replicate what has happened the last 2 days. Theoretically it shouldn't be possible, but believe me when I say that is what happened. I , however, cannot explain why.

Try swopping the live and neutral in that diagram, and insert the dead fault between the "new" neutral and earth.
And remember to bridge the earth and neutral on the supply side too (your East London (and Eskom) special condition ).

Alternatively, if the test button is between the supply side of the phase conductor and the load side of the neutral conductor relative to the earth leakage sensor, and the neutral is bridged on both sides (via the earth), the test fault bypasses the earth leakage sensor completely (or at least not enough current through the neutral conductor coil to cause a trip).

That zero voltage you have between earth and neutral plays some interesting tricks.

[Derlyn]

Brilliant Dave.

You are quite correct.

If the test button is connected to the live on supply side, and neutral on the load side of the elr then no current is going to flow through either sensing coil if there's a neutral earth fault because there's an earth bypass.

Now the next move is to open and reverse engineer an Onesto earth leakage to check the connections inside. I suspect that the test button is as you suggested and not as per the diagram I posted. If that is the case, I would suggest that that's a design fault on the unit itself considering the fact that neutrals are earthed on the supply.

Good observation there, Dave.

[Dave A]

No need for destruction testing. Just swop the poles and test that way