Electronically controlled LED lamps glowing when off. 

I don't understand electronics at all, but i find myself coming back to watch more. Maybe it's the soothing voice and gentle accent, but i've spent hours watching these videos

Can I suggest a video where you experiment with a couple of solutions to this problem? I know you suggested some things to try but it would be neat to see them in action and I think it would get a bunch of views.

I come here for electronics, not art class. You can draw shitty schematics all day, as long as you explain what they do :)

One of those quack power saving devices (PFC capacitor) connected in parallel with the lamp would probably stop the flickering/glowing, without taking much power. Hey, they may actually be useful!

7:40 >capacitor (draws resistor) and resistor (draws capacitor) Well played sir.

Thanks for the warning about SS relays leaking enough juice to bite.

Clive it would be really great if you could do a follow up where you test potential methods of fixing the issue, since it is so common now with household dimmers and LED lamps. Would be interested to see for each potential fix -- snubber network, just a high value resistor, etc. -- the current draw when "off", the waste heat generated, etc. Thank you!

I've got to thank you Clive, I put in a ceiling fan recently and the LED bulb would always stay on at a low level after switching it off and I thought it was a fault with the fan. I swapped the bulb for a halogen one (seems they don't sell incandescent bulbs anymore) and it works perfectly. This video saved me the effort of taking the whole thing out and replacing it for no reason so thank you!

Our old house had a light fitting in our bedroom that had a small leakage current. I remember walking up one night to the CFL bulb in the fitting flashing. It wasn't constant flashing, just a really bright flash every few minutes, then a few quick flashes every so often. It was a bit freaky until I figured out what was going on. It was a bit like those stupid ghost hunting TV shows :-)

I had a CFL in my bedroom, standard UK house wiring, no relay, no dimmer, and it would flash about once every 5 minutes when switched off. Thanks for explaining why it was happening! 👍

I have that glow problem with my X10 controlled wall switch controlling 5 kitchen spots. Initially I installed fluorescent and now Led bulbs. Intuitively I installed one incandescent with the fluorescents and continued when I swapped them for the LEDs. Both types needed the incandescent. Now I know why the problem occurred. Thanks for the enlightenment.

If that's a knock-off of a Fotek, would that make it a Fauxtek?

WOW, Thanks ! I just spent 2 hours trying to figure out why my SSR was showing "almost" full voltage across the Output side, when NO input was present..... I thought I wired it wrong...and you showed me what I had an assumption about - SSR are NOT "Old Fashioned" coil relays that make a NICE noise when opening and closing!

I'm so glad you put this up. It explains why some of the LED lights I put into my truck interior lights never completely go off.. couldn't use them. Can't figure a way to fix it even with your helpful suggestion.. Take Care

Adding a plug-in transformer in parallel to the lamp (via a 1 to 3 outlet tap) seems to fix the issue without hooking anything up to the transformer output. A common old+heavy 5-12v DC "wallwort" is best; experiment to find the least vampiric one for your setup. They fix LED bulb issues with SSRs, dusk-to-dawn sensors, motion detectors, lit switches, etc.

Thanks for your experience and clarification, most low cost SSR and dimmers will display same quirks. I wanted to energize a PWM fan array using the USB voltage to activate the solid state relay. Big failure, fans would just keep running at fractional speed with the relay control voltage removed. Two solutions found: Livolo active phantom load, or a PTC implementation.

Thank you Clive, this has been very very helpful. I have three DMX controlled power junction boxes. Each box has four three-pin-plugs. The boxes, and each plug therein is controlled by a separate DMX controller, which consists of twelve on/off buttons, (so I can run a maximum four boxes of four plugs each.). I've been using them to run a number of LED lights such as Chauvet Swarms, Mini Kinta, Mushroom, spots, and a number of others, for dances and parties. I have never been able to figure out why many of them keep flashing when they were meant to be off, yet work fine when turned on. I had to revert to using normal electric multi plug junction boxes and switching them on manually. I just tried experimenting with a double adaptor, with one outlet going to an LED light and the other to a standard 60W lamp. There was no flashing with both plugged in, but when I unplugged the lamp, the LED started to flash. Once again, thank you. Now I've got to figure out a way of putting a load on to each plug that's the equivalent of a 40~60 watt bulb.

Home made SS relays with quadracs ("snubberless triacs") for the win! Relay or dimmer applications won't ever notice the missing quadrant, and the leakage is essentially zero.

My patio has a pair of remote control track light fixtures, each with 3 lamps. I replaced all with cheap LED bulbs, and they'd come on okay, but when off would strobe weakly. I replaced them with even cheaper but dimmable LED bulbs, and they work perfectly.

you answered a long time question! I have and have used for years x10 dimmer modules for remote control and computer timers. now with regular light bulbs nothing but with CFL or led they all flicker. now it scared the £@#$ out of me at 3 am. but I am so used to it now I have a night light! thanks.

Possible solution: Put an X2 capacitator in parallel with the lamp as a bypass for the leaking current. NOT for dimmed lamps! Phase cutting produces harmonics. These higher frequencies will lead to dangerously high currents throug the capacitator when the lamp ist powered on. If a dimmed LED lamp glows or flickers when turned off, use a "Kompensator 6596" from Busch-Jaeger. This is just a PTC in a casing. Bypassing leaking current to neutral, it will get hot an highliy resistive when the lamp is switched on.

If your going for the one tungsten lamp as shunt for LED circuit concept, you can actually buy decorative large filament lamps. While technically these are much less efficient, they are not designed to light a room, so total current draw is often less, and they have a basically unlimited service life, which is cool given if it does fail, all your LED's will go back to either flashing randomly or glowing when ever the lights are turned off.

I think the last solution was probably the best. include a higher power rated bulb or device in the sequence with a higher power requirement to shunt the leakage to a device that acts as a threshold to the system. Good video.

Found this out recently when I was using a solid state relay to control a neon transformer. Even when it was supposedly off, there was enough current going through the transformer to produce very tiny sparks when the output wires were a fraction of a millimetre apart. Luckily it didn't catch me by surprise, because I was also using the Cliff QuickTest and its neon indicator was also well lit even when the SSR was supposedly off!

I guess you could just put a capacitor between live and nautral to fix this and save 30kW while you're at it.

That was interesting, thanks! I think with the continuous spread of LED lamps, we will have to adapt our electronics to them, so change the whole design of dimmers/solid state relays and so on. It's often like that when you change one part without changing the others.

We had this happening with a normal light switch, but it would flash every ~10 minutes. I am guessing it was the ac coupling in the lightswitch and the cabeling (capacitive effects) that caused this. amazing how little power these lamps actually need to give out light!

had that problem on a CFL, after it was one for a while. every few minutes it flashed very slightly. annoying when youre laying in bed and see this trough your peripheral vision...

For tungsten lamps it is possible to use two 15W bulbs in series as a resistive load so the power usage is not that high and it takes decades for the bulbs to go eol.

There is always a leakage voltage on an SSR .It will allow low voltage low current devices to to stay illuminated .We will switch heater elements through and use a manual motor starter or contractor to disconnect the output going to heaters .Especially at 600V 3 phase heater banks used in commercial ovens .Croydom, Opto,or Carlo Gavazzi they all have the same issues

thanks for giving me info on the! I bought a 100 pack of the clear wago connectors. these look awesome!! thanks guys.

I've been hooked on your videos for the past month now and I don't know why....maybe its your intense knowledge on electronics or maybe it's your interesting accent. either way, i don't plan on stopping watching your videos!!

Hey Clive, big up from Finland!

I have noticed that glowing effect in a couple of led lights hooked up to a dimmer in my house, and thought it was the lights having the problem. Now I know it's the dimmer's snubber network causing the issue :P

Here in Norway we mostly use 2, or 3-fase 230v, and in certain conditions on single pole lights switches theese LED fixtures will continue to glow dim even when off. Only explanation I could find was capacitive leakage through ground... At night the LED lights are bright enough to navigate in "complete" darkness.

the flashing is actually pretty cool for an outside lamp

With big dimmer packs it's often not just the snubber network. Most dimmer packs are used for tungsten lamps, and as you most likely know, to prolong the life of the lamps, there's a pre-heat setting on the dimmer packs. This delivers just enough power to the filaments to heat up, but not enough to actually light up. When you use LED based lamps on these lines, they wil light up, because they need so little energy. That's likely what the problem was with the disney dimmer packs :p

It's true... Every day is indeed a school day.... For many, many years, I always assumed that the leakage current was coming through the triac / SCR (Delete as applicable). I had no idea that it was actually the snubber network....

Cool, a channel worth watching!, thanks man! Loving those dulcet Scottish tones and the information and ideas you provide in your vids, keep it up, you got me interested in electronics again! Regards and greetings from south east UK! :)

Even the stray capacitance in the wiring over here makes some LED's with capacitive droppers light up.

I ran into this problem recently after replacing all light switches in our house with the ones that have the gas-based light indicator(that turns on when you turn off the light). One of the lights in the house is an LED light fixture and when the switch is off, it blinks rather aggressively and frequently(like once every 2 seconds or so). I've been looking up solutions to the problem and I did get four different suggestions: 1. Connect a resistor and a capacitor in series with the light fixture 2. Connect just a resistor in series with the light fixture 3. Connect just a capacitor in series with the light fixture 4. Disconnect the light indicator(duh, but I bought the ones with the indicator on purpose and I'd rather have it work :D ) :D I think I'm going to grab some resistors and some capacitors with different values(suggested resistor values were up to 1 W with resistance between 100k and 1M ohm and for the capacitor either 220nF/400V or 100nF 275VAC) and see if I can resolve the problem.

Best shitty drawings made with CliveCAD© 2016

I knew someone that thought they were getting free light, so they connected a few to make it just bright enough to walk about. I just couldn't get it through to him that he was not getting it for free, He was in the countryside so everything was pitch black, so even the small leakage current would light up things just enough to see to get around. It was so funny he thought he was getting very dim light for nothing, about as dim as he was. Strange thing though, give him an engine and he was brilliant. But electric and him did not mix, he even stuck with car electrics. But he could get more out of an engine than anyone else I knew. He even got more out of a motorcycle engine than the manufacturers could. You just have to laugh sometimes!

Clive, could you please make a video about driving triacs for use as switches and dimmers, and what typical practical circuits look like? Huge undimmed fan of your videos!

Back in 2008 I fitted the then 1.3w GU10 LED bulbs now here is the puzzle for you. I also was able to see the slight glow on these LED Bulbs the switch was a standard light switch, GU10 has no earth point I never Figured a valid reason of how the LEDs were doing this. I checked that the switched wire was live the bulb only connected to neutral with old-fashioned 5A wire fuses at the fuse box disconnecting the live wire from the switch - no change standard 1.5mm flat lighting cable white with red and black conductor wires My only thought was the some kind of eddy current /transformer effect from the steel wall conduit (1950s house that been re-wired using the old conduit that once had the rubber flat cable in)

Clive could you do one of your great videos on building the snubber network to stop the flickering on the LED light and test it. Thanks

Good video. Thanks for the detailed explanation. This reminds me of when using a CFL light and with a LED light plugged into to a socket wired to a old use Dimmer replaced wall switch, the CFL and LED lamps would flash even if the dimmer push-to-be-off was engaged. So irritating at night, the dimmer had to be replaced with the normal on/off switch. - I wonder if a pull-down to ground resistor (of some high value to not load down normal operation) between the 'snubber' capacitor and resistor would prevent the cap from charging and discharging in cycles. Yeah, its not worth the bother to modify the Solid State Relay but still in principle, I wonder if that would solve the issue.

incandescent dimmers have the same or similar issue, they usually list their minimum load as 40w or so. Even a 4 fixture with led bulbs wont meet this.

Great video thanks Clive

Love the Wago connectors but prefer the 221s, smaller and easier to use. PS: Thanks for your posted videos.

3:55 - Sheesh...I almost choked on the hard candy I had in my mouth when he drew that resistor like some kind of weird fish...LOL Also, when he mentions Disney Paris with the "...clumps of tungsten lamps hanging out of the box...", all I could think was, "I guess Disney doesn't live by the _Jurassic Park_ slogan: 'Spare No Expense." LOL

I did actually the same with tungsten lamps to a dimmer rack once. But in this case I didnt had enouhg hanging lights so had to find all kind of desktop lights that night. You could guess how funny that must have been looking at a fast switching part.

You get the same if you use one of those switches with a small neon light for a CFL

Thank you for the detailed explanation! I was starting to belive that there are ghosts in my computers LEDs 0.o

ghost load? never heard any one say that. Ive heard Phantom load. you learn something new everyday

All triacs have a residual current around 7 mA. So if you dont have snubber, you can get the same effect.

Surprisingly I haven't had this happen yet, but then, I still use mostly CFLs, the only LEDs in this house are the LED filament bulbs and one lone candle-bulb style LED COB down in the living room, kind of sad that they don't put the anti-flicker/glow components in place, but every 0.000000047% of a penny helps I guess... :P

I remember our kitchen light on a dimmer system did that years ago,it would spaz out once in a while. My dad was convinced it was a live neutral so he invested in a tester only to find out that it wasnt it. He did change the dimmer to a different type and that stopped the ocassional flash. It usually was associated to something like a lamp being switched on in the next room on the same circuit. We would freak my sister out,telling her it was our dead aunt trying to communicate something. lol.

Those wago lever nuts are fantastic

This must be the reason my outdoor halogen light would just flash with a big LED replacement; there must be a triac/snubber in the PIR detector unit. I'll have to get the ladders out and have it all down again!

Interesting stuff. Thanks for the explanation. :)

I put a new fitting in my bathroom, and populated it with LED bulbs ... the leakage through the switch means it glows at low level when off not noticeable in the day but obvious at night I've decided its a useful feature enough light for a pee in the middle of the night after all at least now the 'wasted power' is light not just heating a filament

I love these videos, I feel smarter every time I watch them. Thank you for making me feel smart!

So the Impedance for this RC network is ZRC = 26.52601 kΩ. So in a 120 volt circuit @ 60 Hz the current is 4.5 MA.

I use an LED night light plugged in to a regular household timer, to let my boys know when they're allowed to wake up on weekend mornings. The night light would glow at night and scare them a bit. I imagine is the same problem?

For my elderly mother I wired in a motion sensing switch (built in the wall to replace a normal toggle wall switch, you understand?) and it did this when not triggered on. Let me be clear -- it did this with an incandescent load ...Would this happen for the same reason? Of course with the Incandescent bulb you did not see a flicker, it just constantly glowed when you wanted to sleep. (it had an adjustable timer for the ON time and a pot for the light and adjusting neither helped.) I thought I was saving money & liked your like playing about with lights & being ahead of the times but in the end it just cost $30 extra to begin with plus loads of my time. Plus I figured it must have leaked like 5-10 watts of power. It was made by a well known lighting and switch maker, too, I forget the name (its pry in one of my electric tool boxes somewhere in the basement) I really want to buy another but it seems like a waste of money ... next time I will probably just try an automated home hub as I've noticed there are ones that do similar things. But it was annoying as all get out. Makes me reluctanct to ever buy another one -- if I ever run across the thing I'll try to edit this to add the name because now its bothering me I can't remember that name.

I could use a PTC resistor to solve the problem?

Good job. I like your explanation style and follow you. But I have a suggestion for you. I see all of your videos, you use single side of your note book. Can you use both sides of the paper? If answer is yes then I will be thankful to you. Please don't mind.

I had an interesting issue when the switch was off (for led ceiling light) but the other switch on the same socket was on (for wall lights) the ceiling light would flash. if both were off then no issue. turns out that voltage was being induced in the neutral (apparently) due to the adjacent circuits and bank charging the capacitor. I had to get rid of the wall lights all together

I have another Problem... I have a LED Strip that just baaaarely glows blue (not even evenly on all LEDs) when it's supposed to be off... Any idea on what could've failed?

Two way switching gives a similar effect here in UK, flickering/dim LED's.

So i have what may be a doggy question. Is it possible to wire a PTC into the light circuit and have it create a load? I have a friend with a large number of recessed ceiling lights in his house ( 45 ) on 4 different 3 way digital dimmers and he wants to convert to LED but unless he leaves one incandescent in each circuit the whole lot all glow when there off. So my half brain idea was to install a 5 watt 80C 110v PTC heater, one on each circuit mounted to a heat sink inside the cans plugged into the circuit to create a slight dummy load. What do you think.....they generate less heat then the 150w max rating that the metal can it rated for was my thought...

I tried putting a 47K 1/2 Watt resistor across the Live and Neutral wires, going to a 240V led bulb. After a few minutes the resistor got to around 48 centigrade, and was too hot to touch. The resistor did not breakdown, even when I extended the on time to 30 minutes. The resistor stopped the led glowing when the relay was off. I also tried connecting a wall wart type power supply across the Live and Neutral going to the 240V led bulb. That also took care of the led glow. So, am I wasting energy, using these SSRS and curing the problem with 'heaters'? Looks like it is more sensible to use mechanical relays for these problem 240V led applications.

Surprised you didn't try to make the extra shunt to show us how / if it would work.

Hi Clive. I have this problem with a light in our stairwell. It has a dual switch so that it can be switched on or off from either the top or bottom of the stairs. Very annoying and i really want to fix this without wasting power. I have no dimmer and no relay in this circuit. Just a switch at top and bottom of the stairs.

Helpful drawing ever

Interesting... way back , found there was enough leakage from the wire to wire capacitance in a 2 way stair light circuit to cause atrocious interference from a CFL when switched off - not enough power to light or flicker, but it wiped out AM radio.

I have some CFLs which I fitted in my garden and one of them flashes when off - not always the same one, and not as quickly as the Asda light in the video. But I know why they do this - the outdoor switch that I wired them to has a neon indicator across the switch that comes on when the switch is off, leaking some current to the CFLs. The neon in turn only lights because of the load through the CFLs.

I had a clapper and I guess it put a small current through the plug to check if something was connected because it caused the CFL bulb in my lamp to blink when it was off and caused my DS charger to behave strangely

I have a weird LED bulb that continues to glow even when unplugged from the socket for hours. It's not always a steady glow, sometimes it will flicker, occasionally glowing bright enough to light up a room for a few moments. Either there is a really big capacitor in it, or it's picking up electricity from some wireless source, like how a florescent bulb lights up near transmission power lines.

+bigclivedotcom I got a good laugh out of the sudden deterioration in quality of your drawings :D

So in what scenario would you use an SSR? Are they build into home lighting circuitry? Clive mentioned set lighting...

excellent video as usual :-)

In my case, it's not a solid state relay, but a completely physical mechanical switch. Heck, even if I completely disconnect the switching wire to the lamp (and leave the live wire connected) and leave the switch off, I measure anywhere between 25V and 40V on the mains with no lamp inserted. However, when a lamp is inserted, I still measure about 5V on the mains. Switch on, and everything's fine and dandy at 230V. I don't have the guts to measure the current, as I'm not sure if it's ok to short out the wires, as I'm not sure where this power is coming from, and why I'm able to measure anything on a wire that is not even connected to anything... So, if a dimmer or a solid state relay isn't the problem and the lamp is switch is completely mechanical, why does this problem still happen? It woudn't be induction, would it? Induction wouldn't cause quite that much voltage, would it?

love this stuff

I found that the cheap SSR would make Led and CFL glow, but not the good high quality SSR like the Crydom.

Lutron makes a synthetic load that mimics a 40W load whilst using only 10W.

I have this problem with cheap LED's I bought on Ebay, they were $1 each in boxes in 10 a few years ago, they work fine, but glow while off, Name brand LED globes bought locally (phillips etc... 5-6 bucks each) Do not do it, Do they have extra circuitry to avoid the problem? or do they just use enough power (6watts instead of 3 like the ebay globes) that they don't glow from the leakage?

I enjoyed your idiots guide to components, because I'm an idiot haha, please do some more Clive, you are a very good teacher for electronics. I think the old BBC should give you a twice weekly programme, they might need a wee warning in the start up sequence, language and shocking jokes :)

Probably the simplest solution is to use a mechanical relay for these applications. You only see this problem on super-low current circuits, meaning ones where a mechanical relay will last indefinitely. There's no lifespan advantage to solid-state in those cases.

Whats the difference in using a fly-back diode or the RC in series used as a snubber?

Couldn't you shunt the leaking voltage off to a infrared LED? It would be intermittently lit / glow, but you wouldn't really be able to see it. Though, I guess you'd have to be careful in case the circuitry produced a strobe in such a way as to turn on/off your television.

This issue seems to be more common with LED lamps of low wattage such as 2.8 watts. I have found that 2.8 watt LED lamp poses this problem whilst a lamp of 5 watt does not. Further lamps of same wattage but of different companies behave differently. Why is it so?

Wouldn't it be better for your energy bill and the environment to not have any leakage current?

Good explanation sir :)

Ever thought of getting one of those Self-Balancing board/Swegway/Hoverboard thingys and dissect them? Maybe get a cheap broken one on Ebay?

I ran into the same thing when I installed one of those lighted "nightlight" style switches in my hallway to control a CFL. There was enough current coming through the light in the switch that every 1.2 seconds the CFL would flash ever so dimly. Took me 3 days to figure out what was going on.

I have a timer for my outside light. Its one of those pole things with 3 LED lamps in it. It flickers from the leakage through the SSR. Mildly annoying, but I could stop it by screwing in a standard lamp if it bothered me enough.

I use a small 7 watt night light on my flashy blinky led bulbs with the motion detector wall switch. The led dimmers don't seem to have this problem.

Wago clips, good love them