Understanding faults/readings

[gazdkw82]

A while back a had an issue with a commercial install I had done for a family members unit

Long story short, 5x50w led floods controlled by 4 robus PIR detectors. Also a 3 way switch for option to turn off, perm on and PIR sensors.

I tested the circuit prior to any work and all the readings where good. I came across an issue where the RCD FCU that I had used on the perimiter trunking kept tripping. Unfortunatly I ran out of time on the original install.

I went back just before xmas to finish a few snags on other works and to look at this fault.

Disconnected the load from the JB and all my works tested fine. Good Zs and I/R over range.

Tested the circuit back to the board and gor a dead short between all conductors. Checked the board and there was 3 outgoing lives from the MCB. I should have confirmed this before I originally installed anything but all the figures were good so I thought it was safe to proceed.

I confirmed the line, neutral and cpc for the lights I have installed. The others I could not confirm so put them in a little JB next to the DB with a label.

The problem I have is I could not I/R test with the others disconnected because there was 20v on the circuit and my tester would not test. As soon as you reconnect the other 2 outgoing line and neutrals the 20v disappears and I can test but get a dead short. I tried line and neutral to earth and that was a dead short also.

I dont understand what was going on. I have left it for over the xmas period as they are closed however I'm still scratching my head.

Any help would be really appreciated.

 

[cliffed]

Sounds like borrowed Neutral somewhere along the line.

 

[gazdkw82]

I suspected the same. When I disc the conductors from the DB the I/R readings were over range so I know the 3 conductors from the JB to the DB are fine.

Disc the other 2 lines and neutrals and I get dead short on the 3 conductors at the JB, disc them from the DB and the tester wont test the 3 from the JB. Hope that makes sense

 

[cliffed]

Yes more investigation will be required,finding which cables are what,good luck keep us informed

 

[gazdkw82]

Yes more investigation will be required,finding which cables are what,good luck keep us informed

Not sure what else I can do. I cant trace the 2 other feeds as the premises is too vast.

 

[marconi]

What instrument are you using to measure a 'dead short'? Is it the same one you are using to measure IR?

If you use an Approved Voltage Tester what voltage readings do you see between an L wire from the MCB and their associated N and E ( ie: LN, LE, NE) and then between all the permutations of these 3Ls, 3Ns and 3Es? Any voltage between a pair of conductors means it is not a dead short between them.

 

[moose]

Are you able to take a new feed to test what you have done. My point for help.

The LED sensors (PIR) floods try to switch the switch side rather than the supply side. I have had this very same thing with triggering sensors.

Good luck.

 

[gazdkw82]

What instrument are you using to measure a 'dead short'? Is it the same one you are using to measure IR?

If you use an Approved Voltage Tester what voltage readings do you see between an L wire from the MCB and their associated N and E ( ie: LN, LE, NE) and then between all the permutations of these 3Ls, 3Ns and 3Es? Any voltage between a pair of conductors means it is not a dead short between them.

Yes i agree. This is why I'm confused. Theoretically a dead short should operate the mcb, unless I'm reading through something.

I'm using a metrel MI3000 for I/R testing.

I cant remember the exact voltage at the JB but it was within nominal voltage. Nothing unusual about the voltage.

I cant test the other line and neutrals because I dont know where they go. Like I said earlier, when disconnected the voltage on the JB supply is 20v and 5v on the others. Connect all back up and 0v on all

 

[Lucien Nunes]

This all sounds pretty normal. The only evidence of a problem is the low IR L-E and N-E on one of the other two branches of the circuit.

Firstly, please can you state a resistance reading rather than 'dead short.' That term means copper touching copper somewhere in the circuit, so typically less than an ohm and in any case low enough to immediately trip the MCB as you say. However, a reading of 0.00 on the megohms scale of an insulation tester could easily be 5kΩ, or even 10kΩ, i.e. ten thousand times higher than a dead short, because the smallest non-zero reading and scale resolution is 10kΩ. That represents a small load of about 5 watts, it won't trip an MCB and isn't even guaranteed to trip an RCD. So we need to be clear about what reading is being given on what instrument, to differentiate between 0.00 on a continuity scale (dead short) versus 0.00 on an IR test scale (small load / very poor insulation.)

The stray voltage on the isolated cables is likely to be capacitive pickup from other live circuits if they share the same trunking, a.k.a ghost voltage. The voltage you read has no particular significance because the cables are floating, i.e. not held at any particular potential by being connected to a supply or to earth. They will take up some arbitrary voltage because a tiny current is flowing into them capacitively from the live circuits. It's a nuisance that the tester doesn't like it but fairly typical; one often has to isolate other nearby cables to get rid of the ghost voltage so that the tester will play ball.

The other two branches of that circuit might have load L-N, more capacitance to earth and little to live circuits, and/or excessive leakage or a fault to earth. When de-energised but connected to the branch feeding your lights, they pass to earth the leakage that's causing the ghost voltage, hence the voltage disappears.

So it all adds up but there might be a fault on one of those other branches.

 

[gazdkw82]

You are exactly right, the reading I got was >0.00 megaohms L/N to cpc. I didnt investigate further as to how long the reading actually was.

The odd thing is that the low reading is only when the other 2 lines are connected, disconnect all 3 lines and the line from the JB is over range on the meter.

The ghost voltage is only prominent on the conductors at the JB when the others are disconnected. So basically I'm in a loop, keep all 3 lines connected at the mcb and get 0.00 megahoms on the meter, disconnect the 2 unknown lines at the mcb and get 20v at the JB and the meter wont test.

The only way I could test the conductors from the JB was to disconnect the line, neutral and cpc from the board. It appeared to be as soon as the neutral from the JB was connected to the neutral bar the readings went as low as the meter would read.

On a side note, I passed 2391 last year. It has to be said that even gaining that qualification doesnt mean you know everything. Experience counts for so much. I passed fair and square and with a good % but sometimes I feel like a fruad when I'm bamboozled by these kind of faults. I feel like I should know what and how to deal with this.

Lucien, your level and depth of knowledge Is inspiring.

 

[Lucien Nunes]

as soon as the neutral from the JB was connected to the neutral bar the readings went as low as the meter would read.

Erm, perhaps I'm missing something but that's what you would expect, isn't it? I'm inferring that the DB was not isolated at source because something must have been live to create the ghost voltage, In which case clearly anything you connect to the neutral bar will be pulled down close to earth. As far as an IR test is concerned, the neutrals would appear to be solidly earthed and give a 0.0MΩ reading. If the DB is 3-phase, the neutral is likely unswitched and so it would remain solidly earthed even with the main switch off.

If the DB main switch isolates the supply N, when testing one circuit's IR with its neutral in the bar, you'll still be reading the N-E insulation of the complete installation including all connected equipment. If your circuit has any load on, then the L-E insulation will read the same because the load will provide a path L-N through which the tester can see N-E.

Again it is useful to consider the different orders of magnitude of resistance typical of different devices and situations. For example, during an IR test, a circuit with a 1kW 230V heating element of 53Ω connected will give a reading of 0.00MΩ L-N, so if the N is connected to the supply N, all three readings L-N, N-E and L-E will be 0.00MΩ. Of course 53Ω is not really a short, it's just a resistance 200 times lower than the lowest indication the IR tester can give on the megohm scale.

 

[gazdkw82]

Erm, perhaps I'm missing something but that's what you would expect, isn't it? I'm inferring that the DB was not isolated at source because something must have been live to create the ghost voltage, In which case clearly anything you connect to the neutral bar will be pulled down close to earth. As far as an IR test is concerned, the neutrals would appear to be solidly earthed and give a 0.0MΩ reading. If the DB is 3-phase, the neutral is likely unswitched and so it would remain solidly earthed even with the main switch off.

If the DB main switch isolates the supply N, when testing one circuit's IR with its neutral in the bar, you'll still be reading the N-E insulation of the complete installation including all connected equipment. If your circuit has any load on, then the L-E insulation will read the same because the load will provide a path L-N through which the tester can see N-E.

Again it is useful to consider the different orders of magnitude of resistance typical of different devices and situations. For example, during an IR test, a circuit with a 1kW 230V heating element of 53Ω connected will give a reading of 0.00MΩ L-N, so if the N is connected to the supply N, all three readings L-N, N-E and L-E will be 0.00MΩ. Of course 53Ω is not really a short, it's just a resistance 200 times lower than the lowest indication the IR tester can give on the megohm scale.

Really useful lucien.

Thankyou