Apparently, even low reverse voltage can kill a LED.

Flashlight Modding and DIY Parts
1

Another lesson to be learned, I guess…

I wasn’t very careful when soldering leads to this LED, and accidentally mixed up polarity.
Connected it to 4A CC source, and got no light - after a bit of confused investigation, I found out that I connected it in reverse, and apparently, this kills the LED!
:frowning:

Pics of what it looked like:

Does anyone here knows REAL reverse voltage limit for LEDs? I don’t think the power supply I used could go to voltage higher than 5V or so, and that’s what reverse voltage specification in XM-L’s datasheet is…

2

It looks more like some strands from the + lead made contact with the copper base, shorting until it melted. An LED is a diode, and acts electrically like any diode. A small reverse voltage shouldn't kill it.

3

the max reverse voltage for the xm-l2 are 5v

4

I didn't think it would be that low.

5

I haven’t tested how much is too much but 4V is in the safe range and it happened two times that I reflowed a LED wrong and plugged it on a power supply with some volts to test it.
Gladly the xml2s have a + mark directly on the led so it is easier to spot.

I guess if it is a cc device it will increase current until it reaches the cc…

6

That solder ball next to the LED might be the culprit.

7

That solder ball is in parallel with the + pad.

None the less the leads are soldered in reverse, black for positive, it should be the other way around.

Gotta try to feed some reverse voltage gradually to some old XM-L and see how it fares.

8

Using a constant current supply on a reversed diode does sound like a way to kill it. Maybe the supply was able to supply more than 5 V for maybe a ms or less. The current might damage it directly or heat the spot where it first broke down enough to make it become permanently conducting, though it could not overheat much of it.
Large silicon control rectifiers need to be protected from too much current when they first turn on, as that would damage the point that opened first before the rest of the diode could open.

9
I very slowly fed reverse voltage to a XM-L up to 5.000V, the current was limited to 0.010A. The LED is fine. I didn't go higher than 5.000V reverse voltage.
I raised the current limit to 4000mA but that that did not change anything, fast turn on, turn off from the switch, didn't either.
Now another power supply could have some transients at turn on or whatever, running after the cheapest thing maybe will get you some spikes for some miliseconds. But this is just one XM-L LED (first generation).
10

Looking at the excess solder on both the + pad and thermal pad I wonder if you have bridged the + pad to the thermal.

I’d remove the LED, clean up the excess solder and reflow.

11

I always use a 3v tester ( 2xAA ) to check + and neg before soldering , then use a marker pen to mark one side ...

This way mistakes are avoided , but you learn from mistakes , the more you learn , the more mistakes you have made ...

Havent made any mistakes , then you havent learned anything