Voltage on light circuit when off

[gazdkw82]

Friend of mine has phoned. He needs my help because his lights in the hallway keep flashing dimly whilst in the off position.

He has measured nominal voltage when the lights are turned on and 50v when the lights are turned off.

I dont have much information to work with yet but thought I would ask for a heads up on how to approach this. Iv not had a fault like this before.

The offending lights are via a 3 gang 2 way switch. Could this be induced voltage? If so, how would I confirm this and how can I remedy it?

He has said he changed the lamps for LED and they flash even more obvious now.

I will obviously check connections and test Zs, probably I/R switch live to neutral/earth etc...

I thought I'd just ask for a heads up on how to approach it.

My friend is an electronics wiz so he has some knowledge on the basics. I would imagine he bas tried the obvious

 

[Lucien Nunes]

I meant, to display 50V when on the wrong setting. But thinking about it again, I recall some MFTs giving a warning that the circuit is live with something like '>50V' when resistance is selected. Perhaps some multimeters do that too (none of mine do, not that I am in the habit of trying to test the resistance of a voltage source) but presumably only when a voltage is actually present?

 

[Ed Evans]

I meant, to display 50V when on the wrong setting. But thinking about it again, I recall some MFTs giving a warning that the circuit is live with something like '>50V' when resistance is selected. Perhaps some multimeters do that too (none of mine do) but presumably only when a voltage is present?

When I assess the AM2 candidates doing their I&T part it is very common that they get a 50V reading due to wrong setting.

 

[DPG]

When I assess the AM2 candidates doing their I&T part it is very common that they get a 50V reading due to wrong setting.

MFTs, not multimeters though surely.

 

[Lucien Nunes]

And presumably most MFTs display a similar '>50V' message for excess touch voltage on a loop impedance test.

 

[static zap]

I used 2 different instruments.

Metrel mi300 multi function tester and a fluke t5-600 meter

.... multi-meters have 10M input impedance...

Sounds like a project for me with my hall strappers !
My MFT , ( has 430k Ohm input on VAC-rms ..from memory)
( Will observe With DMM and MFT in parallel )

 

[gazdkw82]

So iv managed to find some time to crudely draw out the wiring.

You will have to forgive my poor knowledge on lighting circuits. For some people it's easy, for me it's always been a headache. Plus I hardly ever work on lighting wiring.

So, I think what we have here is 2 circuits with switch drops from both upstairs and downstairs lighting, via the red and black at each 2 gang 2 way.

The bit I cant fathom is, why are the colors for the intermediate red and black?

Have I missed something? Is the red and black in the 2 ways not a switch/perm feed?

[automerge]1590052495[/automerge]

some more photos.

may help

 

[Lucien Nunes]

From your sketch, it looks like perfectly normal 'conversion' 2-way with the strappers in 3C+E, where someone has later broken into the runs of strappers and extended them to a new intermediate position using T+E. Since only L1 & L2 need to go via the intermediate (com just carries on through the break-in point as before) they used T+E not 3C+E.

I would say it's a scenario in which the capacitive leakage is likely to be quite high. Lots of cable runs with lots of PL and SL cores near each other.

 

[gazdkw82]

I think your right.

If you drop one of the blue strappers out, the flashing stops but obviously you loose some functionality between switches.

So snubbers are the way forward?

 

[Lucien Nunes]

Yes, but just to be sure that the leakage is not (at least partly) due to damaged cables, I would do some quick IR checks especially between PL and SL including between the two switch groups. My method for that would be to do a normal test L & N to E first with the switches in an 'on' permutation, then:

• Switch one switch over to create an 'off' permutation
• Disconnect PL and SL in both roses, earth both PL and SL in the upstairs rose and SL in the downstairs.
• IR test the downstairs PL to earth, obviously if it shows zero resistance the switches are back in an 'on' permutation so flip one over.
• Switch over both 2-ways leaving the intermediate as-is and test again, to ensure all strappers are tested against common.
• Earth the downstairs PL, un-earth the upstairs PL.
• Repeat steps 1-3 at the upstairs rose.

If you can't get to the roses you can do it all from the 2-ways by disconnecting the switch drops but leaving the strappers in L1 & L2. However, the switch drops are not then included and there is more of a risk of something getting disconnected and missed. It takes longer to describe than it will to do. Basically by earthing all conductors except one PL or SL you are testing the insulation on that relative to everything else of interest. If all results are good then crack on with snubber loads.

 

[gazdkw82]

Yes, but just to be sure that the leakage is not (at least partly) due to damaged cables, I would do some quick IR checks especially between PL and SL including between the two switch groups. My method for that would be to do a normal test L & N to E first with the switches in an 'on' permutation, then:

• Switch one switch over to create an 'off' permutation
• Disconnect PL and SL in both roses, earth both PL and SL in the upstairs rose and SL in the downstairs.
• IR test the downstairs PL to earth, obviously if it shows zero resistance the switches are back in an 'on' permutation so flip one over.
• Switch over both 2-ways leaving the intermediate as-is and test again, to ensure all strappers are tested against common.
• Earth the downstairs PL, un-earth the upstairs PL.
• Repeat steps 1-3 at the upstairs rose.

If you can't get to the roses you can do it all from the 2-ways by disconnecting the switch drops but leaving the strappers in L1 & L2. However, the switch drops are not then included and there is more of a risk of something getting disconnected and missed. It takes longer to describe than it will to do. Basically by earthing all conductors except one PL or SL you are testing the insulation on that relative to everything else of interest. If all results are good then crack on with snubber loads.

I cant get to the Rose's. The 2d fittings only have L/N/E coming through the ceiling so I'd imagine the Rose's are in the void somewhere.

I'll order some snubbers then. Is there any specific type? Rating? Spec?

 

[gazdkw82]

Just about to order snubber from cpc.

How are they connected into the circuit? Across live loop and neutral at the rose for upstairs and the same for downstairs?

 

[Lucien Nunes]

In parallel with the lamps, at any convenient point in the circuit.

 

[gazdkw82]

In parallel with the lamps, at any convenient point in the circuit.

So if 3 lamps are effected, that's 3 snubbers to be fitted?

 

[Lucien Nunes]

One per switched line, it's bleeding off the leakage from SL to N instead of it passing through the lamp(s).

 

[gazdkw82]

So a snubber at every effected fitting between SL and N