You know those LED replacement tubes for fluorescent fixtures? The things that I HATE? Well, Cree has recalled about 700,000 of them because they can melt. cpsc.gov link

If you currently use linear fluorescent lighting, keep using it. If you use LED lighting, throw it away and install fluorescent!

According to the description they are meant to run on the output of fluorescent ballasts

I guess some lamp had a bad connection in some component of its internal driver, and instead of clamping down the ballast voltage it let it go high, the rest is as expected on the OCV of an IS ballast....

It was about time until a recall of LED lamps or fixtures would be invoked. This proves how all LED lighting products are so unreliable, and won't last their 50,000 hours rated life.

Those LED tubes are design to run on 600v of an IS ballast? Most of the retrofit tubes I've seen involve cutting out the middle mans (which in this case is the ballast) and then wiring the existing sockets to the mains voltage.

This proves how all LED lighting products are so unreliable

this proves only one product may be defective.
For example: an company recall an tap valve because it may leak: according to your sentence, all tap valves are unreliable.

I understand a lot of this users here hate leds, but please don't be hater like children.

Those LED tubes are design to run on 600v of an IS ballast? Most of the retrofit tubes I've seen involve cutting out the middle mans (which in this case is the ballast) and then wiring the existing sockets to the mains voltage.

I read somewhere (possibly on Mike Holt Nec forum ?) that bypassing the ballast violates the UL listing of the lantern. If thats true, then there is actually no "approved" way of using direct mains LED tubes at all....(allmost) And that would make sense to why the ballast-output LED tubes are needed. I dont know if its true though

Allmost : I seen 4ft T8 lanterns that come from the factory ballastless with 240 straight to 2 pins in 1 end socket and no wiring whatsoever to the other, those come bundled with LED tubes that require direct mains through 2 pins on 1 end of the tube, (Im trying to imagine the bang that will happen when the LED tube dies and somebody tries putting in a 36W T8 - unballasted 240 across one cathode)

I agree with dor here. LEDisease is a monstrosity and it should be stopped immediately. This is yet another advantage to the side of anti LEDiseasers here.

The fault is one point in our advantage, but throwing statements like that (one recall --> all are bad) only does discredit us completely

The whole idea of LED's replacing the fluorescents in the form of direct retrofit is really crazy - tons of technical issues and only very little benefit, if even any.

And using of fluorescent socket without ballasts is really mad.

Don't understand why with new fixture they just do not use plain LED panels plus separate ballast for them. Way easier to design, manage the heat, so way cheaper and more reliable than those retrofits.

I have quite hard time understanding the persistent tendency to still design new fixtures for incandescents, when it should be clear they will in most cases the incandescents would be never used and it then yields to all the unnecessary troubles with all the retrofit designs (although compare to the fluorescent retrofits, with the "incandescent format" at least the supply is rather fixed specification).


And regadring the recalls:
The fact they do recall the products dy themselve means, they really want to deliver at least some quality level.
If you want to make any new designs, you will have, sooner or later, face some design errors sneaking after all the design validation. That is the fact and was even before the mankind. Doing it completely without any error is just not possible at all.
It is only a matter of how seriously you treat your customers: If you just deny them, but that means just lying to the customers. Short term it may hide the problem, but after few cases the customers would just look for cheaper source.
Or you are sincere and just admit them. Short time it may not bring that good PR, but long time it makes you way more trustworthy, so allow you to hold higher selling prices, so maintain the profit.

Don't remember the barnd of LED retrofit flourescent bulbs installed at work.They run off the line voltage.The ballasts got sent to the recycler.Not all of the lights were converted-the converted ones are for "test".Guess if they work out without going out in flames or not lasting long-we will see-then the other lights get converted.The fixtures are old-from the late fifties and early 60's.Many of the ballasts were PCB.-they did still work fine.Sometimes-most of the time if something works-don't mess with it!

Medved, i agree with you about recalling an product: if you care about customers, you will recall defective products.
I remembered when i bought my first led lamp, after 2 days, an internal screw come loose, i replaced lamp with a new and, after 2 hours, glass broke.
I contacted the producer because hazard of this lamp, but it didn't seemed curious about my experience.
So, i totally banned this producer from my house.

The ones that I have seen are marked that they are only for use on F32T8 instant start ballasts, meaning that they treat the supply as a 265mA AC current source and only rectify it and possibly filter it.  The only electronics the lamp would contain in this case would be a Graetz bridge of ultrafast or high potential rated Schottky diodes and possibly some smoothing caps.  This might be due to something like using the lamps on a different ballast that puts too much current through the lamp, although since all of the Cree fluorescent replacement lamps that I have seen were of Chinese manufacture some kind of shoddy workmanship is not out of the question.

With some designs, an open joint somewhere can end up very bad

Here, its quite possible that the series string of LEDs is the only means to clamps down the ballast output voltage (which is ok as long as its made failsafe in some way). For the LED tube to be of any actual energy saving, its voltage must be lower than that of the fluorescent, so several times lower then the ballast OCV. If there is an open in the string, the voltage across it is ballast OCV, so components like rectifier bridge etc can break down, flash over etc. With the constant current output of the ballast, there is no enough current to blow clear the compoinent or a PCB track etc, so itll just stay there and smolder....

I have done experiments with materials and their vulnerability of certain materials to arcing with an instant start ballast and the circuit board will turn into carbon, which is electrically conductive. Even as low as 12 volt will burn the board ad long as it is initially heated, such as by an arcing or burning component.

A while ago i found a fluorescent lantern where ballasts & wiring were fried, and i did some breakdown tests there too - well, the chokes were allready broken down, they just kept arcing on the coil and smoking when i connected power. I experimented with the wiring too. I will eventually upload the pictures

I managed to draw >10mm long "filaments" in the wire isolation with 240V (current limited by 40W incandescent lamp in series). Once the "filament" gets going, the voltage drop on it is in the order of 50V

And it was just ordinary PVC wire isolation, the section that was near the overheated chokes (hardened and discolored, but not cracked). Makes you think twice when you see hardened old isolation on some ballast wiring and are tempted to treat it as "still good as long as you dont bend it". No, it is not still good !

(Right now the lantern is allready restored with new wiring & chokes)