I'm getting more messages from people who are receiving shocks from faulty LED lights - especially flood lights.
The problem is caused by internal arcing that occurs in LEDs when they are part of large series strings being driven by high voltage supplies. When the LED fails open circuit it starts to arc and track internally, often causing flickering of the light in the process.
Because the aluminium substrate LED panels have an extremely thin layer of insulation between the copper and aluminium, it is easy to damage it when an LED burns up, especially the newer directly mounted flip-chip LEDs. This results in a conductive path to the aluminium, which may not be grounded.
The thin layer of insulation also causes unusual effects when people touch ungrounded LED lights with the switch on the neutral, and the LEDs glow because the person touching the ungrounded light provides a ground path for capacitively coupled current from the LED panel.
With the high number of improperly grounded LED floodlights coming into the country via direct import or from distributors who have not tested their imported products properly, the risk of shock from outdoor lights (and some indoor lights) is increasing. This introduces a few secondary hazards. A metal fence or structure that the light is mounted to may become live, any shock from a light at height could result in a fall and the effects of rectified DC current leakage may affect the ability of some RCD/GFCI devices to trip, and may also prevent tripping when AC leakage occurs from another fault/shock.
I'd recommend that extra precautions are taken to fully isolate faulty metal cased LED lights before working on them, and if there is any doubt about the power source then insulated gloves should be worn and work should only be done in dry weather.
I'd also recommend that the ground integrity of lights is properly tested, and lights with flimsy copper coated aluminium flex have it replaced with something more appropriate if possible.
It's best to source new lights from prominent suppliers in your country that have a reputation to uphold. That automatically excludes most eBay and amazon sellers.
The circuitry on this panel was unusual. It was either an attempt to implement discrete intensity toggling, or to vary the current regulation according to the position in the sinewave.
In this instance the separate LED packages themselves had burned clear before the substrate had been significantly damaged. A 500v insulation test showed it was intact.
If you enjoy these videos you can help support the channel with a dollar for coffee, cookies and random gadgets for disassembly at:-
http://www.bigclive.com/coffee.htm
This also keeps the channel independent of YouTube's advertising algorithms allowing it to be a bit more dangerous and naughty.
#ElectronicsCreators

This circuit board was sent for our perusal a while ago by michael he'd had it in a light and it failed, and it did that usual thing - that the leds all crack and burn out. But it's quite an interesting circuit board and i'll show you the enlarged form of this, since i have not had to take anything apart, it's all ready for photography. So, let's zoom in on this and we'll take a look at the circuitry. We have the incoming supply.

We have two 10 ohm resistors to limit the inrush current. We have a tiny little surface, mount metal oxide varistable, it says vr1, then a bridge rectifier converts the ac into dc and then the option and pads for another metal oxide varistor and then the leds are actually current limited by one two: three uh bp5136h drivers - and This is a typical circuit used with these, but it's not using a typical circuit. Normally you'd have the leds divided into four sections and as the voltage of the sine wave rose up and down, it would initially light the first section and once the current threshold was reached, then like the second and then the third. So as the voltage went up and down, it would like progressive numbers of leds, and it keeps a strange if you look at it through a camera, it sometimes gives us a flickering effect.

I can't do that with this one because uh it's dead, that's why i can't do it. But, interestingly, here for the final stage, they've actually got two 2k resistors 4k in total across that it's quite unusual and they've done that separately to each and every chip. But then commoned all these parallel connections up also, this is a resistor that sets the current by the current sense. So the current flows through the leds goes through the current sense, resistor and uh.

It will regulate the overall current in the circuit, but uh. There is some extra circuitry in this one and it's based around a couple of transistors here that seem to switch in other resistors. The reason there's so many links uh. It looks a busy circuit board, it's not as busy as it looks.

There are tons of links because these chips are effectively commoned with uh three separate circuits driving these leds. But there's a couple of transistors here and i without actually trying the circuit board because it can't because it's dead, uh, i'm going to make a wild guess. Given the circuitry around that it's quite hard to trace out, because it is a aluminium core pcb more about that later, which stops you shining light from the back. But it also has the white solder resist, which makes it very hard tracing the tracks out, particularly when it's a bit grubby and close together like this, but i think these transistors are effectively act as a toggle and might be allowing that function.

You can turn the light on dolphin or change different intensities, not really sure, but what we're interested in right now is what's going on here with these burnt leds, because this is so common in these things, and this is why uh the fact that there's an aluminum Substrate is a bit of an issue so, as happens, you can see damage in other leds here, as happens leds they've baked them now. Are these the hard ones? Are these the soft gel or where are they soft gel? That's gone hard. Now it's pretty hard, uh phosphor loaded sort of resin. Maybe it was gel, i'm not really sure it's hard now.
What can happen is cracks can happen in that and the thermal expansion and contraction and the cycling basically causes a the circuitry that dies inside the bonds to snap off, and you end up with an open circuit when you get an open circuit, because these operate with Quite a high voltage and the open circuit is that high voltage across it you end up this burning a tracking arc, and this is a really major issue with the aluminium substrates pcbs and the reason for that. I shall show you this. This is a standard piece of circuit board material, it's a solid piece of fiberglass and then it's got a thin shim of copper on the surface and to make the circuit board. They photographically transfer an image onto this and then they develop it and then etch off with ferric chloride or copper chloride, and that removes the unwanted copper and just leaves the tracks where they're needed and then there's a lot of other processes.

Drilling and printing. The silk screen and soda resistant things like that, but with the aluminium core pcbs, it's slightly different, the main bulk of the core is a piece of aluminium and the reason for this is for heat dissipation, but on that aluminum then, is a super wafer thin piece Of the fiberglass and then the copper, but the downside of this is that quite often when these fail, i don't know if this one's done it they don't all do it. Let me grab a meter and we'll try. Although, to be honest, i should be testing this at high voltage to get an accurate result here, but if i scrunch my leads through there, i might possibly find.

I don't know that the uh that it's perforated right through to the um, oh very crispy, right through to the uh aluminium as i've found in other ones, and when that happens, i think i need to do a high voltage test on this because uh it will Have broken down the fiberglass underneath, but it's not it's not testing through with a low voltage test, but when it does burn through and there is a fairly common problem, you end up with a dead light and it starts conducting onto the back plate and because uh, Some of these lights are not earthed properly. The wire there's an earth wire into the light, but it's not actually connected to anything. It's just floating thin air. I don't know why they do this, it's very common for the grey imports but uh.

The other issue is this bridge rectifier here, because the current leakage to the aluminium plates in the back or aluminum, if you so prefer, will have a dc bias, and this is where the rcds that would theoretically pick up that leakage um, because this will be current. Limiting so it won't, it won't just go bang and trip the circuit breaker. There will be current limiting through the circuitry here, but uh. The risk is that, even if there's an rcd that you'd think would normally pick that up or gfci the majority of them used in place.
America and britain are type ac circuit, breaker rcds, which are really only designed to detect a sinusoidal leakage like say, for instance, basically the power outside here before it actually reaches here like if the cable got pinched or watched get into the junction box. But when it goes through this, you have a situation where you need a specialist type of rcd to pick that up and germany is way ahead of the pack for ages. They've not allowed ac circuit breakers because they obviously saw the electronics, the progress of electronic stuff happening and all their houses are protected by type, a or b circuit breakers, which can detect the choppy dc in these or smooth dc from from, say something like a car Charger or in the case of the high frequency power supplies they can detect the noise, the high frequency switching output, which can often be defeating. It can often defeat standard circuit breakers.

So this is an issue. It means that you really have to ground these flood lights and make sure that there is connection between the earth wire and the case, give it a good decent test, ideally a pat test. I current that any continuity test is going to be better than nothing in this particular situation um, and it means that really, as time goes on, maybe britain and america are dragging their heels and they need to update and start enforcing the use that all new rcds And distribution boards need to actually be the modern, slightly more sophisticated type that could detect all the other fault scenarios other than just like damaged wiring, so um. That is really the way things are going.

I've come across so many of these that they have leaked onto the case, and people are reporting that they're getting shocks off uh fittings when they go and check why they're not working and they touch them and get a zap because uh it is leaking onto the Aluminium and unfortunately, because either they've not earthed it or grounded it, or it's got that that missing wire. It provides an alternative path of current and that's through them. So uh, yes, maybe the uh iet in the uk, which is our electrical sort of regulatory body, the ones that actually makes up the regulations. Maybe they should actually look at enforcing the progress towards type b circuit breakers, uh rcd, gfci type circuit breakers to detect.

The progressive number of faults like this are going to be increasingly happening in the future.

8 thoughts on “LED shock hazards are increasing”
  1. Avataaar/Circle Created with python_avatars Miguel DeMarchena says:

    (mi idioma ingles no es muy bueno asi que por favor use google translator) yo sustituiria la fuente conmutada de ese articulo por una fuente capacitiva calculando una reactancia que saque a los leds de su zona de estres prolongando la vida de los leds hasta a mas de 20 aรฑos

  2. Avataaar/Circle Created with python_avatars Ethereal Rose says:

    I don't see the issue. You should use a switch to shutdown power to the light before working on it. Risk of shock Is zero and you don't need a gfci/rcd for that. Of course an egc would help maybe but if it is still appropriately current limiting then an egc may do nothing as It acts like a giant resistor. Simple solution is to flip your switch before working on it. Using a multimeter to verify also helps. Electricity is still sent back to the grid as AC so I don't see how a DC gfci would be useful.

    Of course try not to buy cheap Chinese crap (all redundant I know)

  3. Avataaar/Circle Created with python_avatars sportster davidson says:

    This is what happens when you let china make everything , they have low standards so thats what we get , there used to be a time when quality and safty came first not profit ! Are you happy you saved dollar now ? Go spend more on a new light , lets see 19.99 for a cheep china light or 35 for a light made in my coumtry , i will spend a little more and get a better made one , support your own country and maybe things will have higher standard's. When we eccept lower quality we are at fault !

  4. Avataaar/Circle Created with python_avatars Gary Kindt says:

    The moral of this story is write your congressperson and tell he or she that things are coming in from China that are out of spec with US safety requirements.

    The other moral (or is it a morsel in this case) is that whenever possible, use battery driven DC fed devices that are not connected to the AC line. With LED lamps that should be a no brainer. I would love to wire up my house to feed DC to DC driven devices. All fed from a battery bank that one recharges at night when the AC power rates are lower.

  5. Avataaar/Circle Created with python_avatars J J says:

    I'm not really sure how there getting shocked seeing as how the housing is plastic 99.999 percent of the time especially from China what are they spitting on the fingers and putting them in the socket truly I don't get it ok yes the aluminum plate may short or fault but that will not transfer thru the plastic housing unless plastic has all of a sudden become conductive and it forgot to tell the universe

  6. Avataaar/Circle Created with python_avatars exgenica says:

    It seems to be getting more difficult to get meaningful current data on causes of yearly deaths, injuries, and hospitalizations of people in the U.S.
    And, I hope someone can find better data because following CDC OSHA and a few others these specific data were not well presented. Every "source" led to another URLled another link and so on. It was a mess and lot of them were ANCIENT data.

    However, the current numbers MIGHT be something like this…

    500 = workplace and home deaths by outright electrocution and many more from burns, nerve and muscle damage, and chronic severe pain, and suicide by those whose lives were ruined wherein the "cause of death" wasn't recorded as electrocution.

    There are many thousands of devastating electrocution injuries and hospitalizations. I've seen claims of anywhere from 100,000 to well over 200,000 people per year injured with tens of thousands ending up in hospitals with major external and internal trauma including amputations and brain damage..

    Since these lamps are increasingly in the workplace AND homes, I would bet all too often with counterfeit safety stickers, I just went ahead and smushed the data I did find (some of it 10 years old).

  7. Avataaar/Circle Created with python_avatars Robert Herberg says:

    The latest thing in our shop is the florescent tube replacement led tubes that arc and burn through the PCB to the point of bisecting the board itself, which is supporting the weight of the bulb itself.. Gravity always wins, so now we have zip ties on all of our lights. Just in case.

  8. Avataaar/Circle Created with python_avatars Lee Winters says:

    Not only that there making fitting and luminaries with none replacement lamps you end up replacing the whole fitting you can't get replacement parts from these unbranded and unknown manufacturers what ever happend the THORN and PHILIPS they have been out priced by a load of cheap 2 year wonder crap what ever money you save in electricity consumption you end up paying more for replacement you need fittings now that have a standard replacement LED lamps I wonder how many fittings have ended up in land fill.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.