In need of a fault finding wizard

[ross0909]

Hi,im in need of some help.Recently completed a 80 bed nursing home,its been tested and the whole installation is electrically sound bar one thing,i got a really puzzling fault on a lighting circuit protected by a b6 60898 .when the circuit is switched off the lights dim(18w pl lamps).after days of testing procedures r1+r2 insulation r and zs all fine ,the voltage is 42v between cpc and line >any help please

 

[IQ Electrical]

It could be a borrowed neutral or a floating earth fault at that sort of voltage, nightmare to find as you're finding now!

 

[ross0909]

hi,thanks for the reply ,some more info for you,there are 8 3p dis boards from the panel board and ive tried eliminating the neutral fault by removing neutral out of the pb and then testing the problem circiut again and it still shows everytime ,could it be a main supplier problem? the mains is about 30m away from substation and its a tnc-s system.the problem circuit is 5 wall lights controlled via photocell and timeclock.simplest circuit in the building typical.many thanks

 

[Des 56]

Try taking a wander lead from the met.then using that to check the continuity of the cpc and the voltage between the live and cpc,and see if there is a voltage between the cpc and the wander lead earth
Cross fingers it may bring up something thats not apparant as yet

 

[nickblake]

Did you do a Ze test as this should show up any faults on the supply neutral faults can give a high Ze reading especially on TNCS as has been said with getting voltages on the cpc its more likley to be a neutral fault , . if there is a fault on the neutral of the supply this can cause alot of damage if the system is three phase as a total loss of the main neutral could cause 400 volts across sockets and light fittings etc

 

[IQ Electrical]

I don't suppose they could be self-contained emergency lights?

Crazy suggestion, I'm scraping the barrel now!

 

[ross0909]

hi des,thanks for reply if it was a main earth problem why only that circuit with the fault out of so many.i have disconnected all the accessories off this circiut and removed the line at board but when i test voltage between the so called dead live and the earth bar it still shows 42v how is this pos.ive also checked all ceiling grid and metal stud for voltage and its all clear.

 

[ross0909]

yes there are em but as i posted ,ive diconnected all accessories off the circuit to prove the circuit and it still shows 42v,i need help fast my heads about to explode after 3 days of this.

 

[Sparky Ninja]

Try taking a wander lead from the met.then using that to check the continuity of the cpc and the voltage between the live and cpc,and see if there is a voltage between the cpc and the wander lead earth
Cross fingers it may bring up something thats not apparant as yet

I would do this as well, but slightly backwards. I would clamp the lead onto the cpc, then at the board measure it to line and isolate individUal circuits till the voltage goes. Then you would have identified the circuit creating the problem.

 

[nickblake]

you could do a Rn R2 test this way you would confirm that the neutral is satisfactory and eliminate that possability

 

[ross0909]

i still dont understand why if its a neutral fault on the system why is it only one circuit with the fault when on that db there is 36 circuits?

 

[nickblake]

it could be a multitude of things , faulty light fitting , short between neutral and earth , but with a neutral fault as well so you end up with a neutral path to earth via a loose connection causing a high resistance which will cause the lights to glow half , so basically you test the circuits fully to eliminate the basics ,R1R2,RnR2 , IR test Zs tests etc , if all test out split the circuit and work on smaller area's till the fault is found

 

[kmcgold]

you could try this
disconnect the controls
check for voltage on the sw wire side of the time/photo cell,
if voltage on the switch wire to lights then turn all circuits off db and check, if voltage still on cable sw of the
next board etc untill no voltage then turn circuits on one by one untill fault this should
give you the circuit the other fault is on

 

[Qedelec]

Just a shot in the dark, but is there any chance of induced voltage from parallel cables?

 

[SimpleSimon]

Could 42v be an induced voltage of some kind?

 

[Sparky Ninja]

What I do know is that if I was the OP I would be thankful for all these ideas, but probably scratching my head with where to start next....

 

[Kamikaze]

This has the hallmark of a damaged cable. Something like a screw through a T &E cable or a nail through the sheathing into a T & E cable. The most sensible thing to do is to disconnect all connections on this circuit and systematically work your way through each leg. Insulation resistance and continuity tests. I know this is time consuming, but the real headache will be when you find it. Replacing the cable, now that everything is probably plastered and decorated.

 

[IQ Electrical]

I'm sure the OP said he'd done IR tests, that would surely have indicated a fault of sufficient magnitude to 'leak' 42 Volts.

 

[Sparky Ninja]

Hi,im in need of some help.Recently completed a 80 bed nursing home,its been tested and the whole installation is electrically sound bar one thing,i got a really puzzling fault on a lighting circuit protected by a b6 60898 .when the circuit is switched off the lights dim(18w pl lamps)after days of testing procedures r1+r2 insulation r and zs all fine ,the voltage is 42v between cpc and line any help please

I'm sure the OP said he'd done IR tests, that would surely have indicated a fault of sufficient magnitude to 'leak' 42 Volts.

I think so too.

I still think the first thing to do is locate the circuit creating the problem via a method like I suggested. Then when the circuit is found, dissect the crap out of it.

 

[bugsy]

I think the problem is more than likely to be a neutal or switch line fault , what cables are installed (skinned , singles , fp. armour?

 

[whats an rb4]

i agree with widdler on this one induced voltages are always a worry but unlikley dont tjhink property wolud be large enough for this to come into play

 

[cr_sparky]

i have had 42v at a light switch before in a old building,there was a neutral (red single core)connected in the line terminal,with it being an em light circuit (3 core) maybe you have got incorrect polarity?

 

[WDMDL]

OK.. you've done IR tests on that circuit...

the problem circuit is 5 wall lights controlled via photocell and timeclock.simplest circuit in the building typical.many thanks

Timeclock AND photocell? Why? Do you need the lights to go off after a certain time of night?

Is the photocell the NEMA socket type?
Check that is specified as a relay switched type.
If not, these need to pass a small amount of current all the time to power the internal circuit. But generally they will say "suitable for filament lamps only" if this is the case.
Unplug the photocell, see what happens

Same with the timeclock, check that it is a relay switched type too. (listen for the relay click as you operate the override)

But if it's induced voltage that's the problem (sounds very unlikely now), and the luminaires don't have PFC correction capacitors, fitting them will generally pull the induced voltage right down.

Simon.

 

[Doug]

If there is still a voltage present when the circuit is switched off at the MCB then perhaps the MCB is faulty.

This has the hallmark of a damaged cable. Something like a screw through a T &E cable or a nail through the sheathing into a T & E cable. The most sensible thing to do is to disconnect all connections on this circuit and systematically work your way through each leg. Insulation resistance and continuity tests. I know this is time consuming, but the real headache will be when you find it. Replacing the cable, now that everything is probably plastered and decorated.

A link between this circuit and another should show on IR tests between circuits. Alternatively, switch off every circuit except this and look for a voltage on another circuit.

 

[the pict]

move the line to another breaker ? MCB's can sneak voltage past the contacts
Pict