Can I direct drive XML with an 18650?

I'd like to use one of the Cree XML LEDs in a few different things. I'm wondering if I can direct drive one using a lithium cell? (18650) If so what kind of current can I expect?

We have driven our Rainier with the Seoul P7 LED with both the
RCR123A and the 17500 protected cells directly, but the voltage
drop across the P7 is higher than the XML. The output capability
of the 18650 is higher than the other two batteries and may damage
a low volt P7. I would not recommend the 18650 and XML, I think
that the protected RCR123A would be okay. Never use an IMR type
in a direct drive mode of operation. A small voltage increase over
the standard data sheet recommendation will deliver a very large
increase in current to the LED. The bond wires in the XML will not
be able to handle the power.

The Luminus SST-50/90 LEDs can handle much higher current loads.

Curt

No.. Too much juice on a 18650 .

H20doctor said:

No.. Too much juice on a 18650 .

Click to expand...

True, I found out the hard way. Actually my XM-L still lights but now only half of the emitter lights. It didn't turn blue or anything and it did this in the first 10 seconds of operation.
The weird thing is that the half dead XM-L is still brighter than a 2.8amp XM-L drop-in that I have in another 18650 host. The beam looks a little funky though.

Direct drive will get you about 25A instantaneous current (calculated).
Granted the cell will never drive that high, but figure it like a dead short, maybe about 10A.


Use a 0.5 ohm 1watt resistor, will give you 1200ma current. (2 X 1ohm 1/2watt in parallel)

3 of the 1ohm's in parallel will get you 1800ma

4 of the 1ohm's will get you 2400ma

All in parallel.

That is assuming 3.4-3.5vf


I just modified a cheap light using 2 cr123's only to use a 17650 only.
Had a 3.5 2watt resistor behind star, and paralleled a 1ohm with it.
Originally drew 715ma withn the 6v and 3.5ohm, now draws about 760 with the
paralleled 3.5ohm and 1ohm (0.777ohm) on 4.1v

I just have to remember to never use the 2 cr123's in it again.
Would be about 3200ma! KABOOM!!!!!

I have done it, but I used really small wires from the driver board to the emitter to limit current. I use protected Trustfire flames and I get about 3.5 amps on a fresh charge and the amperage drops with the voltage which also dims the light output.

My question is why is the half working XM-L still working and pulling 2.3 amps on a 4.1v 18650?

Here is a shot of it through a #10 shaded welding glass.
The camera makes it look even better than it really is.

Most all XML have a low VF. On a copper heat sink I have direct driven them on a li-ion before, but only for a few seconds. The current rocketed to 6, 7, 8, 9 A and I disconnected, no doubt it would have risen much higher.

One simple solution is to get a 7135 based driver board. They are cheap and easy to use, perfect for single li-ion.

It all depends on the quality of your batteries. My cheap Trustfire cells average only about 4 maybe 5 amps max. So with really good heatsinking it actually works. With my 2600 mah Samsung cells salvaged from a laptop battery they'll draw about +7 amps especially as the led warms up it goes up. IMR, forget it, about half the time it'll instaflash.

Sorry to ask a different question but somewhat relevant, Is it safe to direct drive an XM-L with 3x eneloop AAs and whats the expected output. thanks

I would imagine that 3 nimh cells will have a similar result. If you used good heat sinking, scrony wire, low grade cells and let them full rest after charging then probly yes. If you use wire rated to do the job its doing and quality cells like eneloops I am guessing you would see 8 amps in a hurry.

Actually the resistance in the contacts between the battery and electrical connections would limit the current. I've tried it with lower resistance 5000mah sub-c hobby cells and as long as they aren't soldered into a pack it'll limit itself to only about 4 amps max. I built a camping lamp based on this current limiting theory based on the contacts and D-cell nimh batteries.

Adding an appropriate amount of resistance through the ground path can allow direct drive of an XM-L with an 18650. I have done a resistance limiting setup using a single IMR 18650 and amp draw held steady at 2.8A after 2 minutes. My heatsink design is still in a holding pattern so I do not have any data concerning extended runtimes. Worst case scenario is that the die releases "magic smoke" and you have to start over. That said, it is still my opinion that a regulated driver is the better alternative to direct drive.

I'd guess tunneling messed up the part of the emitter closest to where the bond wires contact the current spreader. Past that, on the opposite side of the chip, perhaps the voltage wasn't high enough to break it.

Cool pic.

I think my problem was heatsinking, I built the light engine in an Aleph E-can and there is not much contact for good heat transfer to the outer body. There is also no room for a good driver in there thats why I went with DD. There was another post somewhere were he made a homemade Aleph light engine with a DD XM-L but I don't know if a resistor or so was used or if he had any problem afterwards using it.

Here is that thread.

Another related question about direct drive. I have an SST-50 on a star that I'm planning on putting into a Dorcy 3D M@G-type light. I know it's safe to run on 3 D cells... what about 4 alkaline C cells? It will be well heat-sinked to a sink and to the body.

Safe? Too much juice? Terribly under-driven?

4 alkaline C-cells is too much. You'll kill the led really quick. Even though c-cells are not as high current as d-cells, at 6 volts, they'll push enough current even with their relatively high internal resistance to fry the led.

Star on big copper heatsink 6 Amp is the max so far I witnessed 3 times. Keep it to 5 at very most.

As long as the discharge rate that the 18650s are capable of are not sufficient to kill it, I don't see why not... I made a direct drive XM-L T6 that I ran off of a blue Ultrafire cell that only maxed out at about 3A so it was cool... If I had tried to stick an IMR in there, it would have turned angry blue and probably died...