LED light still glows softly after being switched off

[Adsibob]

We have a long LED flexible tube light, made by ilumos, recessed in our bedroom ceiling. After being switched off, there is still a very very faint glow from it. Maybe 1% of its lumens. I googled this fault, and it appears there are several possible causes, from poor wiring, to poor bulbs to poor light switches:

https://ledstore.pro/blog/2023/02/06/why-led-stays-on/#:~:text=Thanks to their high efficiency,to produce a dim light.

 

the light in question can be switched on or off from three different light switches, all using the same brand of light switch, which I believe is this:

 

https://www.corston.com/products/intermediate-toggle-antique-brass?variant=39293017751649&bcndyn=d2lkOjQ4OHxjaWQ6NTY5&gad_id=&utm_campaign=Google_Ads&gclid=Cj0KCQiAgK2qBhCHARIsAGACuzl1FlNR3r1YE3F-6TPccW9ae0fi-GGOtwnMByEOeoQZSo5eXcBeDy4aAoBbEALw_wcB

 

That switch is made in China, and although it was very expensive I suspect the manufacturer focused its efforts on the brass switch plate rather than the internal switch module. 
 

The wiring is all brand new (less than 2 years old) so I’m hoping it’s not that.
 

Can anybody recommend a high quality switch module that I could swap out for the three switches to see if this is the cause? It needs to be a toggle switch, not a rocker. 

 

[JohnMo]

I had a similar issue on a motorcycle once, had to change the brake switch in the end. Was fine with a normal bulb.

[Adsibob]

The mild irritation is that this is in our bedroom. 

[SimonD]

Our new electrician has recommended Varilight for all our led switches. They do sell toggle switches but ours will all be dimmer. I don't have direct experience yet, obviously.

[Bramco]

We have BG Evolve -> https://www.bgelectrical.uk/uk/wiring-devices/evolve   In our case all black but they seem to have some bronze-ish ones.

Doesn't leave a glow when turned off although we will probably have different led strips.

 

They do have one rather annoying feature though - after a power cut, they come back in the 'on' state when power is re-applied!!!

So annoying if there's a power cut at night but even worse if you're on a long holiday.....

 

But apart from that, we're very happy with them - we hate those rotary dimmers and also small fiddly switches.

 

Simon

[Bramco]

1 hour ago, Adsibob said:

The mild irritation is that this is in our bedroom. 

There is a +ve to that - you won't stub your toe against the bedframe while navigating to the loo in the middle of the night  😄

 

Simon

[Adsibob]

38 minutes ago, Bramco said:

There is a +ve to that - you won't stub your toe against the bedframe while navigating to the loo in the middle of the night  😄

 

Simon

it's not even that bright! But me being the perfectionist that I am find it annoying. Weirdly, i just tried the light in that room and the issue didn't occur. I wonder if maybe it only happens after the light has been on for an hour or so, maybe it only "glows" after it has "heated up "(but even that would be surprising given it's an LED).

Edited November 8, 2023 by Adsibob

[Adsibob]

I think I found what the issue is and the potential solution another forum:

 

It's capacitive coupling between the wires between the two switches, this allows a small amount of electric current to make a side-run around the open switch contacts. This small current is enough to cause some CFLs to flash periodically and some LED fixtures to glow dimly. and to trigger non-contact voltage detectors.

It can be cured by adding a lighting capacitor in parallel with the light fixture, providing a path to neutral sufficient to pass most of the ghost current rendering to voltage at the outlet too low to activate the lamp.

 

 Thoughts?

[Mattg4321]

You're on the right tracks. As I understand it, when you've got long runs of cable where a 'live' conductor is run for a distance alongside one that is not (as in 2 or 3 way switching), then a 'ghost' voltage can be induced. Not enough to give you a shock, or to light an incandescent lamp, but it can make some LED's glow.

 

I've come across it a few times, but always when there is 2,3,4 etc way switching in place. Not so much lately though, I think the better manufacturers have got a handle on it now.

 

You need a 'snubber', such as this one from TLC. Connect in parallel with the load. It'll need to go inside an enclosure/light fitting/switch if there's room.

 

https://www.tlc-direct.co.uk/Products/DNCAPLOAD.html

[LnP]

I experience something similar with some external LED lights which are in a 3-way switched circuit. The two outside light switches have small red "locator' lights so you can find them in the dark. I'm wondering if the reason the LEDs glow when the switches are open, is because the the sparkie wired the circuit with no neutral in the switch boxes, and the locator lights in parallel with the switch contacts. So the locator lights would get their supply from the live in the box and the small current powering the locator light would find its way back to neutral via the LED fitting, causing the LED to glow.

 

Would that make sense?

[SteamyTea]

Inductance/capacitance between wires is quite common.

You can think of it as an inefficient transformer.

Uncoil the wires from the iron core, lay the side by side, pop some AC power through one and you will get some power in the other wire.

It is how Marconi got radio signals across the Atlantic.

It is a useful feature of oscillating circuits.

A double pole switch will sort it out.

[Adsibob]

20 hours ago, SteamyTea said:

A double pole switch will sort it out.

Where would the double pole switch go in my three way switching set up? Is it instead of one of the switches I have already or in addition?

[SteamyTea]

34 minutes ago, Adsibob said:

Where would the double pole switch go in my three way switching set up? Is it instead of one of the switches I have already or in addition?

Ah, probably won't get one that does all that.

The point I was making was that if there total disconnection, the problem goes away.

[Adam2]

I had similar but it was the LED bulbs used in a bedroom ceiling. Thought it was my automation boxes causing a small leakage current but was fixed by changing the bulbs - I guess some bulbs have a "feature" that does nothing with any very low level current? Maybe a quicker thing to check?

[Temp]

On 08/11/2023 at 20:35, Mattg4321 said:

You need a 'snubber', such as this one from TLC. Connect in parallel with the load. It'll need to go inside an enclosure/light fitting/switch if there's room.

 

https://www.tlc-direct.co.uk/Products/DNCAPLOAD.html

 

+1 It essentially just a capacitor. You wire it in parallel across any light fitting. Normally only need one even in a room full of down lights. 10 min job just don't pull all the wires out and forget where they go 🙂

 

[Adsibob]

On 10/11/2023 at 09:46, Adam2 said:

I had similar but it was the LED bulbs used in a bedroom ceiling. Thought it was my automation boxes causing a small leakage current but was fixed by changing the bulbs - I guess some bulbs have a "feature" that does nothing with any very low level current? Maybe a quicker thing to check?

This is a pretty long flexible LED tube, but not so flexible that I’m keen to remove it from the recess in our ceiling and experiment. It’s odd, because although it’sa long LED tube, I’m pretty sure it runs off normal voltage, without a transformer.

[SteamyTea]

12 hours ago, Adsibob said:

This is a pretty long flexible LED tube, but not so flexible that I’m keen to remove it from the recess in our ceiling and experiment. It’s odd, because although it’sa long LED tube, I’m pretty sure it runs off normal voltage, without a transformer.

The 'transformer' will be built in.

The inductance/capacitance will be in the mains wiring leading to the light, not in the light itself.

[FuerteStu]

Just curious, do you have pictures of his the lightswitches are wired? 

 

I've seen massive differences in induced voltage when 2 way/intermediate switches are wired with 3 core, with feed and switch going to same plate, and the old school method of running the feed to the first switch and taking the switch live up from the last switch plate. 

 

(which I thought was against regs since 17th Edition because of the induction loop it created) 

[AartWessels]

On 08/11/2023 at 20:21, Adsibob said:

I think I found what the issue is and the potential solution another forum:

 

It's capacitive coupling between the wires between the two switches, this allows a small amount of electric current to make a side-run around the open switch contacts. This small current is enough to cause some CFLs to flash periodically and some LED fixtures to glow dimly. and to trigger non-contact voltage detectors.

It can be cured by adding a lighting capacitor in parallel with the light fixture, providing a path to neutral sufficient to pass most of the ghost current rendering to voltage at the outlet too low to activate the lamp.

 

 Thoughts?

That's correct. 

It is a very common problem with led lights, regardless of quality of wiring, switches and light fixtures. 

[FuerteStu]

On 08/11/2023 at 23:43, LnP said:

I experience something similar with some external LED lights which are in a 3-way switched circuit. The two outside light switches have small red "locator' lights so you can find them in the dark. I'm wondering if the reason the LEDs glow when the switches are open, is because the the sparkie wired the circuit with no neutral in the switch boxes, and the locator lights in parallel with the switch contacts. So the locator lights would get their supply from the live in the box and the small current powering the locator light would find its way back to neutral via the LED fitting, causing the LED to glow.

 

Would that make sense?

Yes, I've seen a string of outside garden lights pulse dimly because if these neons. they are easy to disconnect and throw away though

[Adsibob]

1 hour ago, FuerteStu said:

Just curious, do you have pictures of his the lightswitches are wired? 

 

I've seen massive differences in induced voltage when 2 way/intermediate switches are wired with 3 core, with feed and switch going to same plate, and the old school method of running the feed to the first switch and taking the switch live up from the last switch plate. 

 

(which I thought was against regs since 17th Edition because of the induction loop it created) 

I don’t, but I doubt anything has been done against regs, add the sparky was very good.

[Mattg4321]

On 10/11/2023 at 06:55, Adsibob said:

Where would the double pole switch go in my three way switching set up? Is it instead of one of the switches I have already or in addition?

It won’t work in any circumstances where this is a problem that I can think of. In ‘3 plate’ wiring there is only one pole to switch at the wall anyway (there is no neutral, and in any case this isn’t what is causing the issue).
 

If the switches are ‘fed’ then you won’t get the problem in the first place as there isn’t a permanently fed live/line conductor running alongside a switched live/line within the same cable, like there would be with 3 plate. 
 

Usually with 1 way switching you won’t get this problem as the switch wire cable won’t be long enough to cause an issue. It’s when it’s 2/3/4 way it can do. Just fit the snubber/capacitor and forget about it. Might even be able to do it in the switch if you’re lucky. 

 

1 hour ago, FuerteStu said:

Just curious, do you have pictures of his the lightswitches are wired? 

 

I've seen massive differences in induced voltage when 2 way/intermediate switches are wired with 3 core, with feed and switch going to same plate, and the old school method of running the feed to the first switch and taking the switch live up from the last switch plate. 

 

(which I thought was against regs since 17th Edition because of the induction loop it created) 

That would be news to me. 
 

The worst thing about the ‘old’ way was when people did it incorrectly and linked the downstairs and upstairs lighting circuits by using a line from one and a neutral from the other. Eg. Landing light used line from downstairs hall switch and 2x strappers to landing switch and switch wire up to landing light, using the neutral from upstairs and the line from downstairs. Quite common a few decades ago and something I still have to put right occasionally now. 

[FuerteStu]

1 hour ago, Mattg4321 said:

It won’t work in any circumstances where this is a problem that I can think of. In ‘3 plate’ wiring there is only one pole to switch at the wall anyway (there is no neutral, and in any case this isn’t what is causing the issue).
 

If the switches are ‘fed’ then you won’t get the problem in the first place as there isn’t a permanently fed live/line conductor running alongside a switched live/line within the same cable, like there would be with 3 plate. 
 

Usually with 1 way switching you won’t get this problem as the switch wire cable won’t be long enough to cause an issue. It’s when it’s 2/3/4 way it can do. Just fit the snubber/capacitor and forget about it. Might even be able to do it in the switch if you’re lucky. 

 

That would be news to me. 
 

The worst thing about the ‘old’ way was when people did it incorrectly and linked the downstairs and upstairs lighting circuits by using a line from one and a neutral from the other. Eg. Landing light used line from downstairs hall switch and 2x strappers to landing switch and switch wire up to landing light, using the neutral from upstairs and the line from downstairs. Quite common a few decades ago and something I still have to put right occasionally now. 

The significance is current flow. 

 

In 3 plate switch wires the current flow goes too and from the switch in the same twin. 

 

In modern 2 way it's the same! Goes down to the switch in twin, all the way along the straps to the other switch plate, then back down the common. So the current flow goes back and forth in the same run. 

 

Old school sparks tend to use singles and earth's as a feed to the switch, then a single up from the other end as a switch live. This causes a  loop in current flow rather than it going back along the same run. 

[Mattg4321]

31 minutes ago, FuerteStu said:

The significance is current flow. 

 

In 3 plate switch wires the current flow goes too and from the switch in the same twin. 

 

In modern 2 way it's the same! Goes down to the switch in twin, all the way along the straps to the other switch plate, then back down the common. So the current flow goes back and forth in the same run. 

 

Old school sparks tend to use singles and earth's as a feed to the switch, then a single up from the other end as a switch live. This causes a  loop in current flow rather than it going back along the same run. 

I’ve not seen insulated and sheathed singles used like that in anything installed since about the 70’s. 
 

Either way I’m not sure it’s subject to any ban, but would be interested to see a source if you have one? It’s still a common way of wiring in pvc/steel conduit. 

[Adsibob]

14 hours ago, Mattg4321 said:

Just fit the snubber/capacitor and forget about it.

Will the snubber/capacitor that was linked to earlier in the thread work, regardless of the voltage requirements of my LED?