scary indeed. lighting science leds are not that good. they make the ecosmart leds sold at home depot and these get more negative reviews than the philips and cree leds.

http://wqad.com/2013/03/19/led-light-bulbs-recalled-for-fire-hazard/

What fault is happening there that makes them actually catch fire ?

And how aboout that green product isnt it ? Besides returning and recycling all those lamps, amount of air pollution from 1 fire >>>> amount of air pollution those lamps prevent

It could be some overlooking in the internal design - like the fusible resistor's functionality being compromised by close proximity of other components in the final assembly, not anticipated during the electronic design.
Or the thermal contact between the LED's and heatsink could be not as reliable, causing severe LED overheat, but at the same time failure of the overtemperature protection, as this in fact sense only the heatsink temperature.
Well, these are problems these tight and high temperature devices really are asking for...

thanks for sharing. turns out i have one that was over the sink, on a dimmer, ironically the first one i bought was a dud and failed within seconds
this one has worked fine, its in a metal thermally protected fixture with good airflow and ceramic socket...and is'nt in an area with anything flammable so maybe ill keep using it



ECOsmart par 20 LED

I'd like to see pics of the inside to try to spot the problem

LED lamps that have HF drivers, have an electrolytic capacitor, so it can catch fire if overloaded too much.
Defect in the manufacturing can cause any electronic device to cause fire.
This can also be a false recall.

LED lamps that have HF drivers, have an electrolytic capacitor, so it can catch fire if overloaded too much.
Defect in the manufacturing can cause any electronic device to cause fire.
This can also be a false recall.

The newest ones do not contain the electrolytes on the primary side anymore - there is only a film capacitor.
And on the secondary side become more and more popular to use a polymer electrolyte capacitor - there the electrolyte is a form of plastic, so it become combustible too (compare to the water based electrolytes)...
The reason is the desire to prolong the high temperature operational life, but the cost could be the way more combustible (compare to classic aluminum oxide elco's) nature of those conmponents.

Im thinking more like a film capacitor catching fire, the phosphor dome melting and collapsing on the LED chips overheating them to the point of fire (of the melted dome), switching circuit failure causing the LEDs to run hotter etc

I seen many blown electrolites in SMPSs and didnt yet see a case that capacitor caught fire. Te closest i seen is a monitor where the 400v electrolite blew and set the rectifier bridge on fire, but it didnt catch fire itself (allthough it splashed quite a bit of electrolite around)

I seen many blown electrolites in SMPSs and didnt yet see a case that capacitor caught fire. Te closest i seen is a monitor where the 400v electrolite blew and set the rectifier bridge on fire, but it didnt catch fire itself (allthough it splashed quite a bit of electrolite around)

You are talking about aluminum electrolytic capacitors, which use water based electrolyte, so they are basically not combustible.
But these new electrolytes use polymers instead of the water (similar as LiIon or more close the LiPoly batteries), so the capacitor's core is combustible. So when they overheat and expose the core to the air (if it explode), it may catch fire.
Externally they look very alike the normal aluminum electrolytic, they use the same aluminum can with plastic/rubber seal.

And there is one other aspect of considering, if the thing is a fire hazard or not:
It is the used fuse rating, mainly it's ability to interrupt fault current.
If you feed too much current into a fuse, the thin wire would melt sooner. But that does not mean the fuse interrupted the circuit.

Normally the blown wire draw an arc (yes, a discharge), what is still conductive, so the fuse still conduct the current, it is not yet OFF. In order to break the current, the arc have to extinguish.
And here come the main problem:
The high current may blow the fuse wire sooner, but it form a fatter, more persistent arc. And when the current exceed some limit, the arc would become so persistent, it won't disappear before becoming a fire thread on it's own. This threshold is called "current breaking capability". The regular instrument 20mm long, 5mm in diameter glass fuse have the current breaking capability of 20A. So if the fault current exceed 20A, nobody guarantee this fuse would be able to stop it.

And for the fuse to be functional, the failure current should never exceed the current breaking capability of the used fuse.
As these electronic devices tend to fail into a hard short circuit, it mean the fuse current breaking capability have to be above the mains short circuit current. And here we talk about at least 1500A. That require, in the 20x5mm fuse format, ceramic body and sand fill.

Other option in dealing with this problem is using dome means to limit the fault current, so it can not exceed the breaking capability of the used fusing mechanism. This is the main feature of the "fusible resistors": Beside providing the mechanism to break the circuit, the resistance limit the short circuit current below the current breaking ability. For low power devices, ~10Ohm are used, what mean limiting the current below 23A, so quite easy to handle then. Moreover it make sure, than such faulty lamp become disconnected in time, so wont trip the breaker for the complete circuit, so other lamps on that circuit stay working.

And the problem is, than some purchasers tend to be cheap and order standard resistors in place of the fusible ones, or a small, insufficiently rated glass fuses, both being cheaper than the original fusible resistor. Then the thing become a fire hazard, because such "fusing" does not work.
Well, the safety of all low power (where the main circuit breaker is of too high trip current) electronic devices is quite vulnerable to this type of manufacturing error, it is by far not limited to the CFL's or LED's. And these are the majority of all "fire hazard" recalls I know about...

From computer PSU's i seen (which do use the 20mm glass fuse) usually when a "bang" happens in the circuit it trips the C16 breaker (so currents on the order of >=80-160A). along with the glass fuse, inrush thermistor and current sense resistors rattling in pieces inside the PSU box

For CFLs i see othger problem why this is a problem : Light fixtures may be wired on relatively thin wiring, in long runs (think of a multi way switching system with a couple old and oxidised switches too) which possibly might by itself add enough resistance to prevent the breaker from tripping for a short

In a couple exploding USB charger cases that i checked out, despite them being totally pathetic (a standard 0.125w warm white colored resistor for a fuse, with 0.5mm clearance between its solderings on the othger side), the circuit still "banged" pretty well to break the arc as longer wires going to the plug pins fused. On the ohter hand i seen a CFL protected by fusible resistor that looks like of decent size, that arced internaly and the pressure opened the ballst enclosure - It remained a screw base with half of the casing, from that hanging on 1 wire + 1 fusible resistor the ballast, from it the tube. Now this one totally could have shot out a spark that would land on some papers or clothes....

Beside the computer PSU (as it is of really high power rating device), none of the devices (below 40W) should trip the breaker. Well, nor the incandescent bulb, but they do so quite regularly.

The fusible resistor have to have enough space around, or it should be inside a glass mat sleeve, so the eventual arcing does not jump somewhere else. Indeed, a condition only rarely observed.

But these faults do not make the LED's anything worse than the other products: Incandescents, mobile chargers, CFL's,...

And for the fixture wiring: I think there is a reason for the minimum allowed wiring cross-section. And if that is not observed and something bad happen as a consequence, it is primarily fault of the fixture...

These plug-in power supplies and chargers are potentially even more dangerous when used on a British 32A ring main! It is a surprise to me that these effectively unfused devices are permitted here where the only local protection is likely to be the fuse in a socket strip or multi-way adaptor, assuming there is one. At least most (domestic) lighting circuits have a 6A (or 5A) breaker.

I had a cup boiler explode once, went off with a BIG flash and bang tripping the 32A breaker. The force from that could easily have split open a plastic box leaving the exposed parts still in the socket.


BG

If there is a correct fuse, there is no danger.
The fuseis there to prevent the firein case the converter itself does fail.

The short and open load states onthe output should be handled by the controller without overstressing any component of the converter, so it should pose no danger at all even when (nearly) permanent. This is usually the case even with the cheapest designs - as it is an intrinsic property of the selfoscillating circuits and it is an integral part ofthe design inside of the dedicated IC's for these converters (there are no IC's on the market not haveing such functionality).
Sometimes is the problem with overloading - there is no hard short, but the load curent is way above the rating (but still below the electrical capability of the convertor). But these cases do not occure when the supplies are used correctly (for the specified applications).
And it become a standard for the newest controllers to be able to keep even the maximum overload current very close to the rated output (so 1A rated source limit the current to or shut down already at 1.1..1.2A)

BG you are right and this happens quite a lot
http://club.dx.com/reviews/photos/48615/234846

i seen few cases myself - in one i was called in the lil hours of the night (it was in a military base and i was in the emergency first reaction team that night) to pull out the 2 pins sticking in the socket holes - all what remained, the rest flew off as the charger exploded, just like you described. The incident hapened about a minute after it was plugged in according to the user who caled me

All cheap chargers seem to share very bad design flaws. Thats the flaws i amsdd Medved described (clearance, inappropriate fusing) and more
http://www.lygte-info.dk/info/usbPowerSupplyTest%20UK.html

C32 breaker need about 160-320A to trip immediately, ie short circuit loop on the order of 1 ohm. This can indeed be a big problem as the short circuit may be quite a bit more with those chargers. A single "bang" is unlikely to cause fire (unless there is spark exelled that lands on papers etc) but this cannot be said about continuous heating if the circuit was not opened right away

But this one picture loos more scary to me. What heated here was near down the circuit and whether there was a fuse (or a substitute) or no, it did not open the circuit and let the components down the road keep going
http://club.dx.com/reviews/photos/48615/145407