Lights problem with LED lamps not turning off

[Doomed]

Bungalow with loft conversion, DB 17th edition split load, so all circuits RCD protected.
Corridor has 3 wall lights, 2 on one switch, 2nd on its own switch. Also kitchen lights, lounge and living room on this circuit.
Customer fitted G9 (I think) LED lamps to hall fittings, and they don't fully turn off, always emitting a low glow. Checked voltage L to N and L to E and both read approx 60V. Interestingly the act of putting my fluke multimeter across the terminals to take a reading turned the LED lamps off - capacitor in the fluke?
Checked through the circuit as best as I could around the loft extension - small crawl space and lots of insulation - found several joints all on choc blocks for this circuit, by splitting them up I isolated the circuit to just the hall lights and the lounge lights, the cable then goes under the extension. The fault was still there - isolated the lounge off at the dimmer switch, fault still there.
Tried taking out the LED lamps - voltage still there.
Isolating the MCB removes the voltage.
Can't get to any more of the circuit, so any ideas where the stray voltage is coming from, or can I just fit something like a snubber and mask the problem?

 

[Pat H]

Not sure the 3% VD would be relevant for a 12V feed? The 12V will be dropped even lower in the LED anyway. Looking online the 12v LEDs will run at 10V or less.
So assuming you use 1.5mm cable you have approx. 14A to play with that's over 20 lamps.

 

[Pat H]

One solution would be to drop the 230V locally for distribution. So a feed to say 1st floor lights. Drops to 12V then switched and distributed. Cable runs would be short and you'd still be ensuring all user accessible items were at 12V not 230V.

 

[Bobster]

Not sure the 3% VD would be relevant for a 12V feed? The 12V will be dropped even lower in the LED anyway. Looking online the 12v LEDs will run at 10V or less.
So assuming you use 1.5mm cable you have approx. 14A to play with that's over 20 lamps.

Why wouldn't you think the VD should count here? No difference in designing a 12V circuit to a 230V circuit. Albeit one is ELV.

 

[Taylortwocities]

I'd need to check the regs, but I don't think it makes any distinction between VD tolerances on LV and ELV systems.
Personally, I don't like the idea of having one driver for all the lamps. All the lights go out if/when the driver dies.

 

[Pat H]

Interesting trying to find why 3% was the chosen max volt drop for lighting. Not been able to uncover why that was decided. But I imagine its to allow for older lighting that has no self regulation properties. Like resistive loads. But now we have LED lights that have to convert to constant current internally at sub a few volts variations on the supply voltage are less critical.
Yes of course the current regulations don't allow for it and until revised regulations decide to cater for it it would be outside regs to go that route. Unless you are able and prepared to justify it...

 

[Pat H]

I'd need to check the regs, but I don't think it makes any distinction between VD tolerances on LV and ELV systems.
Personally, I don't like the idea of having one driver for all the lamps. All the lights go out if/when the driver dies.

That's a valid point but easily solved if needed. There are already various supply solutions that offer redundant backups. A second parallel convertor that waits to switch in. The other advantage of 12V distributed would be the ease of adding battery backup for power outages.

 

[Midwest]

Interesting trying to find why 3% was the chosen max volt drop for lighting. QUOTE]

Not forgetting the supply of voltage can be + 10%, - 6% of 230 volts, so your light bulb might be trying to operate at 209 volts?

Not forgetting the supply of voltage can be + 10%, - 6% of 230 volts, so your light bulb might be trying to operate at 209 volts?

 

[Pat H]

209.714V I calculate...

 

[Midwest]

209.714V I calculate...

So did my calculator, but the lamp should be able to function at 209.714 volts, but possibly not at 209 volts

 

[Pat H]

To add an additional variable. Many LED drivers work from 110V to 240V so at 209V its more than happy.

 

[Midwest]

I believe not all drivers are the same. Universal AC input drivers can operate on varying voltage ( e.g. 90v to 360V for worldwide use), other drivers seem to operate on a range of primary input of 220V - 240V for UK use, so they wouldn't be happy, nor for that matter would BS7671.

 

[Pat H]

nor for that matter would BS7671.

What about BS7672 or whatever it will be

 

[Richard Burns]

I think the lighting limit is primarily for discharge lighting that may fail to strike at lower voltages; resistance lighting will "work" at any voltage it will just get dimmer and dimmer.
This is why you get such problems as the OPs because a tungsten filament lamp would have been "working" at the low voltage but not enough power would be supplied to generate a glow but the voltage would be discharged, with the LEDs they can operate on this limited power and glow dimly or flicker depending on the LED circuitry.

 

[Doomed]

Have you checked your main earth rod, and your main bond, are they both in good order ?

Should of said - yes, TNS, good readings, tested RCD's while I was at it, both OK,voltage around 240.

 

[Doomed]

Get one of these
RC Contact Suppressor | Maplin - http://www.maplin.co.uk/p/rc-contact-suppressor-rg22y
and fit it across one of the lamps (where you put your Fluke multimeter).

Ordered some - thank you - will go back on Thursday and fit.