Sorry for posting twice. I accidentally posted without any text, and couldn’t figure out how to edit my post.
I have two Duracell 1,000 lumen LED flashlights that use three “C” batteries (4.5 V in series). I’m pretty sure the surface mounted Cree XML LED is what failed. With my multi-meter in diode mode, it reads short and doesn’t light at all. (I tried my meter on a 12 V car taillight LED and it lights up).
I can’t find a bead 1,000 lumen LED to re-solder, or a replacement COB. (I could just buy new flashlights, but I’m trying to learn electronics repair) The driver board is in the flashlight end cap. The momentary switch (ON/OFF, HI/LOW, STROBE) seems fine, the capacitors all read the same value, and the transistors read the same. The IC has continuity between pin 1 and 4.
Thanks for any advice! Pics below.
We have some XML LEDs in stock here, but I don’t see one that matches yours. If you can provide the specifications (voltage, current, and dimensions), we can check if we have a similar option available.
The LED itself is a surface mount 5 mm x 5 mm (5050) The flashlight uses 3 C batteries, so I think the forward voltage is between 3.6 to 4.5 V. I don’t think the driver steps up the voltage. I suspect the driver primarily allows high and low intensity, and generates the pulse for the strobe mode.
I don’t know how many milliamps it draws at max output. The flashlight is marked 1,000 lumen, and only has the single LED. I’m assuming the 1,000 lumen is the luminous flux (perceived brightness) vs. radiant flux (i.e. total EMR).
This would be my suggestion for a replacement LED.
The board it’s mounted on is most likely a metal-core PCB for heat transfer purposes, which are pesky to solder on using the typical pencil iron for that reason. Hot plate and/or heat gun of some type are likely solutions in an improvised context. The trick is to get enough heat in to do the job, and get it out again before things are cooked to oblivion…
SMD291AX-ND is my preferred solder paste–I wouldn’t even bother with wire-type solder for such things. 243-1055-ND is my preferred tweezer and I like a narrow solder wick for cleanup, again because the fat stuff acts like a heatsink.
The 1000 lumen figure is likely stretching the truth near its elastic limit. Listed flux ratings apply at the listed test current; output scales linearly with drive current to a first approximation with diminishing returns, so if a person takes an LED rated for 300lm @ 0.7A and drives it at 4x that figure, it’s possible one could manage whilst the LED survives.
‘Aint necessarily conclusive; there appears to be a resistor in parallel with the LED for some reason (likely to mess with things) and many meters’ diode settings don’t measure blue LEDs well. Not sure what’s happening with your taillight unit, but getting any visible light out of it with a diode test is odd.
Close enough for full credit. “Luminous” in context implies a measurement weighted according to standard models of human spectral sensitivity. This is a decent wikipedia entry offering more information.
1 Like
Thanks Rick. I saw that the maximum forward voltage on the XM-L2 is about 3 V. IDK what the max V is for the Cree XML, but I’m guessing it’s in a similar range. I have two of these flashlights and I left one intact. They use four C batteries in series. The flashlights have high, low, and strobe modes that cycle with a press of the switch. On high, I measured about 5.8 V across the LED. (The 47 Ohm resistor looks like it’s in parallel.)
It seems that nearly twice the forward V would quickly fry the new LED. The voltage oscillates in strobe mode from about 1.5 to 3 V. Low mode is about 1 V. Now I’m wondering if the LED drivers failed to properly regulate the voltage and fried the LEDs. I tried applying 3 V directly to the LEDs and they didn’t light.
Guess I’ll practice desoldering and reflow on these components.
Here’s some new info for others who may be trying to diagnose one of these or a similar flashlight. I incorrectly stated that the flashlight used 3 C batteries –it uses 4, so closer to 6 V input.
I soldered leads onto the driver board and the LED’s MCPCB, put them on my breadboard and applied 5V. The driver and the LED appeared to be shorted. I swapped out the LED board for a conventional LED indicator light. It lit up but nothing happened when I pushed the switch button on the driver. I was hoping the fresh LED would output hi/low/strobe and turn off.
I’m sure many thousands of these Duracell flashlights were sold, but I can’t find boards that fit. The driver board with button switch goes in the end-cap and is 20mm diameter. The MCPCB with the Cree XML LED is marked 23.5 mm. I de-soldered the original LED and can probably use an XML-2 as suggested, but it would be easier to replace the board with a pre-installed LED.
